This page intentionally left blank World Scientific N E W J E R S E Y • L O N D O N • S I N G A P O R E • B E I J I N G • S H A N G H A I • H O N G K O N G • TA I P E I • C H E N N A I Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE Library of Congress Cataloging-in-Publication Data Stakhov, A. P. (Alexey Petrovich) The mathematics of harmony : from Euclid to contemporary mathematics and computer science / by Alexey Stakhov ; assisted by Scott Olsen. p. cm. -- (Series on knots and everything ; v. 22) Includes bibliographical references and index. ISBN-13: 978-981-277-582-5 (hardcover : alk. paper) ISBN-10: 981-277-582-X (hardcover : alk. paper) 1. Fibonacci numbers. 2. Golden section. 3. Mathematics--History. 4. Computer science. I. Olsen, Scott Anthony II. Title. QA246.5.S73 2009 512.7'2--dc22 2009021159 British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Copyright © 2009 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher. For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher. Printed in Singapore. EH - The Mathematics Harmony.pmd 1 7/17/2009, 9:41 AM Dedication v With deep gratitude to my parents, my darling father, Peter Stakhov, who was killed in 1941 during World War II (19411945), and to my darling mother, Daria Stakh, who passed away in 2001, and to my darling teacher Professor Alexander Volkov, who passed away in 2007. This page intentionally left blank Preface vii Preface Academician Mitropolsky’s commentary on the scientific research of Ukrainian scientist Professor Alexey Stakhov, Doctor of Engineering Sciences I have followed the scientific career of Professor Stakhov for a long time— seemingly since the publication of his first book, Introduction into Algorithmic Measurement Theory (1977), which was presented by Professor Stakhov in 1979 at the scientific seminar of the Mathematics Institute of the Ukrainian Academy of Sciences. I became especially interested in Stakhov’s scientific research after listening to his brilliant speech at a session of the Presidium of the Ukrainian Academy of Sciences in 1989. In his speech, Professor Stakhov reported on scientific and engineering developments in the field of “Fibonacci computers” that were conducted under his scientific supervision at Vinnitsa Technical University. I am very familiar with Stakhov’s scientific works as many of his papers were published in various Ukrainian academic journals at my recommenda tion. In April 1998, I invited Professor Stakhov to report on his scientific re search at a meeting of the Ukrainian Mathematical Society. His lecture pro duced a positive reaction from the members of the society. At the request of Professor Stakhov, I wrote the introduction to his book, Hyperbolic Fibonacci and Lucas Functions, which was published in 2003 in small edition. In recent years, I have been actively corresponding with Professor Stakhov, and we have discussed many new scientific ideas. During these discussions I became very impressed with his qualifications and extensive knowledge in regard to his research in various areas of modern science. In particular, I am impressed by his knowledge in the field of mathematics history. The main feature of Stakhov’s scientific creativity consists of his uncon ventional outlook upon ancient mathematical problems. As an example, I shall begin with my review of his book Introduction into Algorithmic Measurement Theory (1977). This publication rewarded Professor Stakhov with recogni tion in the field of modern theoretical metrology. In this book, Professor Sta khov introduced a new mathematical direction in measurement theory—the Algorithmic Measurement Theory. Alexey Stakhov MATHEMATICS OF HARMONY viii In 1993, I recommended a publication of an innovative paper, prepared by Professor Alexey Stakhov and Ivan Tkachenko, entitled “Fibonacci Hyper bolic Trigonometry,” for publication in the journal Reports of the Ukrainian Academy of Sciences. The paper addressed a new theory of hyperbolic Fibonacci and Lucas functions. This paper demonstrated the uniqueness of Stakhov’s scientific thinking. In fact, the classical hyperbolic functions were widely known and were used as a basis of nonEuclidean geometry developed by Ni kolay Lobachevsky. It is quite peculiar that at the end of 20th century Ukrai nian scientists Stakhov and Tkachenko discovered a new class of the hyper bolic functions based on the Golden Section, Fibonacci and Lucas numbers that has “strategic” importance for the development of modern mathematics and theoretical physics. In 1999, I also recommended Stakhov’s article “A Generalization of the Fibonacci QMatrix”—which was presented by the author in English—to be published in the journal Reports of the Ukrainian Academy of Sciences (1999, Vol. 9). In this article, Professor Stakhov generalized and developed a new theory of the Qmatrix which had been introduced by the American mathe matician Verner Hoggatt—a founder of the FibonacciAssociation. Stakhov introduced a concept of the Qpmatrices (p=0, 1, 2, 3...), which are a new class of square matrices (a number of such matrices is infinite). These matrices are based on socalled Fibonacci pnumbers, which had been discovered by Sta khov while investigating “diagonal sums” of the Pascal triangle. Stakhov dis covered a number of quite unusual properties of the Qpmatrices. In particu lar, he proved that the determinant of the Qpmatrix or any power of that matrix is equal to +1 or 1. It is my firm belief that a theory of Qpmatrices could be recognized as a new fundamental result in the classic matrix theory. In 2004, The Ukrainian Mathematical Journal (Vol. 8), published Stakhov’s article “The Generalized Golden Sections and a New Approach to Geometri cal Definition of Number.” In this article, Professor Stakhov obtained mathe matical results in number theory. The following are worth mentioning: 1. A Generalization of the Golden Section Problem. The essence of this generalization is extremely simple. Let us set a nonnegative inte ger (p=0, 1, 2, 3, ...) and divide a line segment АВ at the point C in the following proportion: CB AB = AC CB p Preface ix We then get the following algebraic equation: xp+1 = xp + 1. The positive roots of this algebraic equation were named the General ized Golden Proportions or the Golden pproportions tp. Let’s ponder upon this result. Within several millennia, since Pythagoras and Plato, man kind widely used the known classical Golden Proportion as some unique number. And at the end of the 20th century, the Ukrainian scientist Stakhov has generalized this result and proved the existence of the in finite number of the Golden Proportions; as all of them have the same right to express Harmony, as well as the classical Golden Proportion. Moreover, Stakhov proved that the golden pproportions τp (1≤τp≤2) represented a new class of irrational numbers, which express some un known mathematical properties of the Pascal triangle. Undoubtedly, such mathematical result has fundamental importance for the develop ment of modern science and mathematics. 2. Codes of the Golden рproportions. Using a concept of the golden рproportion, Stakhov introduced a new definition of real number in the form: A= ∑ a τ , (a ∈ {0,1}) i i p i i He named this sum the “Code of the golden рproportion.” Stakhov proved that this concept, which is an expansion of the wellknown Newton’s definition of real number, could be used for the creation of a new theory for real numbers. Furthermore, he proved that this result could also be used for the creation of new computer arithmetic and new computers—Fibonacci computers. Stakhov not only introduced the idea of Fibonacci computers, but he also organized the engineering projects on the creation of such computer prototypes in the Vinnit sa Polytechnic Institute from 19771995. 65 foreign patents for inventions in the field of Fibonacci computers have been issued by the state patent offices of the United States, Japan, England, France, Germany, Canada, and other coun tries; these patents confirmed the significance of Ukrainian science and of Pro fessor Stakhov’s work in this important computer area. In recent years, the area of Professor Stakhov’s scientific interests has moved more and more towards the area of mathematics. For example, his lecture “The Golden Section and Modern Harmony Mathematics” delivered at the Seventh International Conference on Fibonacci Numbers and their Applications in Graz, Austria in 1996, and then repeated in 1998 at the Ukrainian Mathematical So ciety, established a new trend in Stakhov’s scientific research. This lecture was impressive and it created wide discussion on Stakhov’s new research. Alexey Stakhov MATHEMATICS OF HARMONY x Currently, Professor Stakhov is an actively working scientist who publishes his scientific papers in many internationally recognized journals. Most recently, he has published many fundamental papers in the international journals: Com puters & Mathematics with Applications; The Computer Journal; Chaos, Solitons & Fractals; Visual Mathematics; and others. This fact demonstrates, undoubt edly, tremendous success not only for Professor Stakhov, but also for Ukrainian science. Stakhov’s articles are closing a cycle of his longterm research on the creation of a new direction in mathematics: Mathematics of Harmony. One may wonder what place in the general theory of mathematics this work may have. It seems to me that in the last few centuries as Nikolay Lobachevsky said, “Mathematicians have turned all their attention to the advanced parts of analytics, and have neglected the origins of Mathematics, and are not willing to dig the field that has already been harvested by them and left behind.” As a result, this has created a gap between “Elementary Mathemat ics”—the basis of modern mathematical education—and “Advanced Mathe matics.” In my opinion, the Mathematics of Harmony developed by Profes sor Stakhov fills that gap. Mathematics of Harmony is a huge theoretical contribution to the development of “Elementary Mathematics,” and as such should be considered of great importance for mathematical education. It is imperative to mention that Professor Stakhov focuses his organiza tional work on stimulating research in the field of theory surrounding Fibonacci numbers and the Golden Section; he also assists in spreading knowledge among broad audiences inside the scientific community. In 2003, under Professor Stakhov’s initiative and scientific supervision, the international conference on “Problems of Harmony, Symmetry, and the Golden Section in Nature, Sci ence, and Art” was held. At this conference, Professor Stakhov was elected as President of the International Club of the Golden Section, confirming his of ficial status as leader of a new scientific direction that is actively progressing the modern science. Professor Stakhov proposed the discipline “Mathematics of Harmony and the Golden Section” for the mathematical faculties of pedagogical universities. In essence, this mathematical discipline can be considered the beginning of mathematical education reform—which is based on the principles of Harmony and the Golden Section. It should be noted that such discipline was delivered by Professor Stakhov during 20012002 for the students and faculty of physics and mathematics at Vinnitsa State Pedagogical University. I have no doubts about the usefulness of such discipline for future teachers in mathematics and physics. I believe that Professor Stakhov has the potential to write a textbook Preface xi on this discipline for pedagogical universities, and also a textbook on Mathematics of the Golden Section for secondary schools. It is clear to me that “Mathematics of Harmony,” created by Professor Stakhov, has huge interdisciplinary importance as this mathematical disci pline touches the bases of many sciences, including: mathematics, theoretical physics, and computer science. Stakhov suggested mathematical education reform based on the ideas of Harmony and the Golden Section. This reform opens the doors for the development of mathematical and general education curriculum. It would greatly contribute to the development of the new scien tific outlook based on the principles of Harmony and the Golden Section. Yuri Mitropolsky Doctor of Sciences in Theoretical Mechanics, Professor Academician of the National Academy of Sciences of Ukraine Academician of the Russian Academy of Sciences Honorable Professor: The Mathematics Institute of the National Academy of Sciences of Ukraine EditorinChief of the Ukrainian Mathematical Journal This page intentionally left blank Contents xiii Contents Introduction: Three “Key” Problems of Mathematics on the Stage of its Origin xix Acknowledgements xxxix Credits and Sources xIiii Part I. Classical Golden Mean, Fibonacci Numbers, and Platonic Solids Chapter 1. The Golden Section 1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 1.7. 1.8. 1.9. 1.10. 1.11. 1.12. 1.13. 1.14. 1.15. Geometric Definition of the Golden Section Algebraic Properties of the Golden Mean The Algebraic Equation of the Golden Mean The Golden Rectangles and the Golden Brick Decagon: Connection of the Golden Mean to the Number π The Golden Right Triangle and the Golden Ellipse The Golden Isosceles Triangles and Pentagon The Golden Section and the Mysteries of Egyptian Culture The Golden Section in Greek Culture The Golden Section in Renaissance Art De Divina Proportione by Luca Pacioli A Proportional Scheme of the Golden Section in Architecture The Golden Section in the Art of 19th and 20th Centuries A Formula of Beauty Conclusion 2 7 11 20 24 25 27 32 37 41 46 51 53 56 59 Chapter 2. Fibonacci and Lucas Numbers 2.1. 2.2. 2.3. 2.4. 2.5. Who was Fibonacci? Fibonacci’s Rabbits Numerology and Fibonacci Numbers Variations on Fibonacci Theme Lucas Numbers 60 62 67 72 77 Alexey Stakhov MATHEMATICS OF HARMONY xiv 2.6. Cassini Formula 83 2.7. Pythagorean Triangles and Fibonacci Numbers 85 2.8. Binet Formulas 90 2.9. Fibonacci Rectangle and Fibonacci Spiral 95 2.10. Chemistry by Fibonacci 97 2.11. Symmetry of Nature and the Nature of Symmetry 100 2.12. Omnipresent Phyllotaxis 105 2.13. “Fibonacci Resonances” of the Genetic Code 110 2.14. The Golden Section and Fibonacci Numbers in Music and Cinema 112 2.15. The Music of Poetry 116 2.16. The Problem of Choice: Will Buridan’s Donkey Die? 120 2.17. Elliott Waves 125 2.18. The Outstanding Fibonacci Mathematicians of the 20th Century 129 2.19. Slavic “Golden” Group 132 2.20. Conclusion 136 Chapter 3. Regular Polyhedrons 3.1. 3.2. 3.3. 3.4. 3.5. 3.6. Platonic Solids Archimedean Solids and Starshaped Regular Polyhedra A Mystery of the Egyptian Calendar A DodecahedronIcosahedron Doctrine Johannes Kepler: from “Mysterium” to “Harmony” A Regular Icosahedron as the Main Geometrical Object of Mathematics 3.7. Regular Polyhedra in Nature and Science 3.8. Applications of Regular Polyhedrons in Art 3.9. Application of the Golden Mean in Contemporary Art 3.10. Conclusion 137 144 148 152 154 160 163 172 179 182 Part II. Mathematics of Harmony Chapter 4. Generalizations of Fibonacci Numbers and the Golden Mean 4.1. 4.2. 4.3. 4.4. A Combinatorial Approach to the Harmony of Mathematics Binomial Coefficients and Pascal Triangle The Generalized Fibonacci pNumbers The Generalized Golden pSections 186 189 192 199 Contents xv 4.5. The Generalized Principle of the Golden Section 4.6. A Generalization of Euclid’s Theorem II.11 4.7. The Roots of the Generalized Golden Algebraic Equations 4.8. The Generalized Golden Algebraic Equations of Higher Degrees 4.9. The Generalized Binet Formula for the Fibonacci рNumbers 4.10. The Generalized Lucas pNumbers 4.11. The “Metallic Means Family” by Vera W. de Spinadel 4.12. Gazale Formulas 4.13. Fibonacci and Lucas mNumbers 4.14. On the mExtension of the Fibonacci and Lucas pNumbers 4.15. Structural Harmony of Systems 4.16. Conclusion 202 204 206 212 214 221 227 232 237 241 249 253 Chapter 5. Hyperbolic Fibonacci and Lucas Functions 5.1. 5.2. 5.3. The Simplest Elementary Functions 255 Hyperbolic Functions 259 Hyperbolic Fibonacci and Lucas Functions (StakhovTkachenko’s Definition) 264 5.4. Integration and Differentiation of the Hyperbolic Fibonacci and Lucas Functions and their Main Identities 268 5.5. Symmetric Hyperbolic Fibonacci and Lucas Functions (StakhovRozin Definition) 277 5.6. Recursive Properties of the Symmetric Hyperbolic Fibonacci and Lucas Functions 280 5.7. Hyperbolic Properties of the Symmetric Hyperbolic Fibonacci and Lucas Functions and Formulas for Their Differentiation and Integration 283 5.8. The Golden Shofar 286 5.9. A General Theory of the Hyperbolic Functions 291 5.10. A Puzzle of Phyllotaxis 299 5.11. A Geometric Theory of the Hyperbolic Functions 301 5.12. Bodnar’s Geometry 307 5.13. Conclusion 313 Chapter 6. Fibonacci and Golden Matrices 6.1. 6.2. 6.3. Introduction into Matrix Theory Fibonacci QMatrix Generalized Fibonacci QpMatrices 317 322 326 Alexey Stakhov MATHEMATICS OF HARMONY xvi 6.4. Determinants of the QpMatrices and their Powers 6.5. The “Direct” and “Inverse” Fibonacci Matrices 6.6. Fibonacci GmMatrices 6.7. Fibonacci Qp,mMatrices 6.8. Determinants of the Qp,mMatrices and their Powers 6.9. The Golden QMatrices 6.10. The Golden GmMatrices 6.11. The Golden Genomatrices by Sergey Petoukhov 6.12. Conclusion 330 333 334 340 343 345 348 350 357 Part III. Application in Computer Science Chapter 7. Algorithmic Measurement Theory 7.1. 7.2. 7.3. 7.4. 7.5. 7.6. 7.7. The Role of Measurement in the History of Science 360 Mathematical Measurement Theory 364 Evolution of the Infinity Concept 370 A Constructive Approach to Measurement Theory 377 Mathematical Model of Measurement 382 Classical Measurement Algorithms 385 The Optimal Measurement Algorithms Originating Classical Positional Number Systems 389 7.8. Optimal Measurement Algorithms Based on the Arithmetical Square 392 7.9. Fibonacci Measurement Algorithms 396 7.10. The Main Result of Algorithmic Measurement Theory 401 7.11. Mathematical Theories Isomorphic to Algorithmic Measurement Theory 408 7.12. Conclusion 413 Chapter 8. Fibonacci Computers 8.1. 8.2. 8.3. 8.4. 8.5. 8.6. A History of Computers Basic Stages in the History of Numeral Systems Fibonacci pCodes Minimal Form and Redundancy of the Fibonacci pCode Fibonacci Arithmetic: The Classical Approach Fibonacci Arithmetic: An Original Approach 416 424 429 434 443 449 Contents xvii 8.7. Fibonacci Multiplication and Division 8.8. Hardware Realization of the Fibonacci Processor 8.9. Fibonacci Processor for Noisetolerant Computations 8.10. The Dramatic History of the Fibonacci Computer Project 8.11. Conclusion 455 460 465 470 475 Chapter 9. Codes of the Golden Proportion 9.1. 9.2. Numeral Systems with Irrational Bases 476 Some Mathematical Properties of the Golden pProportion Codes 480 9.3. Conversion of Numbers from Traditional Numeral Systems into the Golden pProportion Codes 484 9.4. Golden Arithmetic 488 9.5. A New Approach to the Geometric Definition of a Number 492 9.6. New Mathematical Properties of Natural Numbers (Z and Dproperties) 497 9.7. The F and LCodes 500 9.8. Numbertheoretical Properties of the Golden pProportion Codes 505 9.9. The Golden Resistor Dividers 511 9.10. Application of the Fibonacci and Golden Proportion Codes to DigitaltoAnalog and AnalogtoDigital Conversion 515 9.11. Conclusion 520 Chapter 10. Ternary MirrorSymmetrical Arithmetic 10.1. 10.2. 10.3. 10.4. 10.5. 10.6. A Ternary Computer “Setun” TernarySymmetrical Numeral System TernarySymmetrical Arithmetic Ternary Logic Ternary MirrorSymmetrical Representation The Range of Number Representation and Redundancy of the Ternary MirrorSymmetrical Numeral System 10.7. MirrorSymmetrical Summation and Subtraction 10.8. MirrorSymmetrical Multiplication and Division 10.9. Typical Devices of Ternary MirrorSymmetrical Processors 10.10. Matrix and Pipeline MirrorSymmetrical Summators 10.11. Ternary MirrorSymmetrical DigittoAnalog Converter 10.12. Conclusion 523 527 530 533 538 544 546 553 557 561 565 567 Alexey Stakhov MATHEMATICS OF HARMONY xviii Chapter 11. A New Coding Theory Based on a Matrix Approach 11.1. A History of Coding Theory 11.2. Nonsingular Matrices 11.3. Fibonacci Encoding/Decoding Method Based upon Matrix Multiplication 11.4. The Main Checking Relations of the Fibonacci Encoding/Decoding Method 11.5. Error Detection and Correction 11.6. Redundancy, Correcting Ability, and the Advantages of the Fibonacci Encoding/Decoding Method 11.7. Matrix Cryptography 11.8. Conclusion 569 579 581 584 588 597 601 613 Epilogue. Dirac’s Principle of Mathematical Beauty and the Mathematics of Harmony: Clarifying the Origins and Development of Mathematics E.1. Introduction 615 E.2. The “Strategic Mistakes” in the Development of Mathematics 620 E.3. Three “Key” Problems of Mathematics and a New Approach to the Mathematics Origins 632 E.4. The Generalized Fibonacci Numbers and the Generalized Golden Proportions 633 E.5. A New Geometric Definition of Number 641 E.6. Fibonacci and “Golden” Matrices 644 E.7. Applications in Computer Science: the “Golden” Information Technology 646 E.8. Fundamental Discoveries of Modern Science Based Upon the Golden Section and “Platonic Solids” 649 E.9. Conclusion 657 References 661 Appendix: Museum of Harmony and Golden Section 675 Index 685 Introduction xix Introduction Algebra and Geometry have one and the same fate. The very slow successes did follow after the fast ones at the begin ning. They left science in a state very far from perfect. It happened, probably, be cause mathematicians paid the main at tention to the higher parts of the Analy sis. They neglected the beginnings and did not wish to develop those fields, which they finished once and left them from behind. Nikolay Lobachevsky Three “Key” Problems of Mathematics on the Stage of its Origin 1. The Main Stages of Mathematics Development W hat is mathematics? What are its origin and history? What distin guishes mathematics from other sciences? What is the subject of math ematical research today? How does mathematics influence the development of other sciences? To answer these questions we refer to the book Mathemat ics in its Historical Development [1], written by the phenomenal Russian math ematician and academician, Andrew Kolmogorov. According to Kolmogor ov’s definition, mathematics is “a science about quantitative relations and spa tial forms of real world.” Kolmogorov writes that “the clear understanding of mathematics, as a spe cial science having its own subject and method, arose for the first time in An Alexey Stakhov MATHEMATICS OF HARMONY xx cient Greece at 65 centuries BC after the accumulation of the big enough actual material.” Kolmogorov points out the following stages in mathematics de velopment: 1. Period of the “Mathematics origin,” which preceded Greek mathe matics. 2. Period of the “Elementary Mathematics.” This period started during 65 centuries BC and ended in the 17th century. The volume of mathematical knowledge obtained up to the beginning of 17th cen tury was, until now, the base of “elementary mathematics”—which is taught at the secondary and high school levels. 3. The “Higher Mathematics” period, started with the use of variables in Descartes’ analytical geometry and the creation of differential and integral calculus. 4. The “Modern Mathematics” period. Lobachevsky’s “imaginary geom etry” is considered the beginning of this period. Lobachevsky’s ge ometry was the beginning of the expansion of the circle of quantita tive relations and spatial forms—which began to be investigated by mathematicians. The development of a similar kind of mathemati cal research gave mathematicians many new important features. 2. A “Count Problem” Discussing the reasons of mathematical occurrence, Kolmogorov specifies two practical problems that stimulated the development of mathematics during its origin: count and measurement. A “count problem” was the first ancient problem of mathematics. It is em phasized [1] that “on the earliest steps of culture development, the count of things led to the creation of the elementary concepts of natural number arith metic. On the base of the developed system of oral notation, written notations arose, whereby different methods of the fulfillment of the four arithmetical operations for natural numbers were gradually developed.” The period that culminated in the origin of mathematics germinated the “key” mathematical discoveries. We are talking about the positional princi ple of numbers representation. It is emphasized in [2] that “the Babylonian sexagesimal numeral system, which arose approximately in 2000 BC, was the first numeral system based on the positional principle.” This discovery under lies all early numeral systems created during the period of mathematics origin Introduction xxi and the period of the elementary mathematics (including decimal and binary systems). It is necessary to note that the positional principle of number representa tion and positional numeral systems (particularly the binary system), which were created in the period of mathematics origin, became one of the “key” ideas of modern computers. In this connection, it is also necessary to remem ber that multiplication and division algorithms, used in modern computers, were created by the ancient Egyptians (the method of doubling) [2]. However, the formation of the natural number’s concept was the main result of arithmetic’s development in the period of mathematics origin. Natu ral numbers are one of the major and fundamental mathematical concepts— without which the existence of mathematics is impossible. For studying the properties of natural numbers, the number theory—one of the fundamental mathematical theories—arose in this ancient period. 3. A “Measurement Problem” Kolmogorov emphasizes in [1], that “the needs of measurement (of quantity of grain, length of road, etc.) had led to the occurrence of the names and des ignations of the elementary fractions and to the development of the methods of the fulfillment of arithmetic operations for fractions.... The measurement of areas and volumes, the needs of the building engineering, and a little bit later the needs of astronomy caused the development of geometry”. Historically, the first “theory of measurement” arose in ancient Egypt. It was the collection of rules, which the Egyptian land surveyors used. As the ancient Greeks testify, geometry—as a “science of Earth measurement”—had originated from these rules. However, a discovery of the “incommensurable line segments” was the “key” discovery in this area. This discovery had been made in the 5th century BC in Pythagoras’ scientific school at the investigation of the ra tio of the diagonal to the side of a square. Pythagoreans proved that this ratio cannot be represented in the form of the ratio of two natural num bers. Such line segments were named incommensurable, and the numbers, which represented similar ratios, were named “irrationals.” A discovery of the “incommensurable line segments” became a turning point in the devel opment of mathematics. Owing to this discovery, the concept of irrational numbers, the second fundamental concept (after natural numbers) came into use in mathematics. Alexey Stakhov MATHEMATICS OF HARMONY xxii For overcoming the first crisis in the bases of mathematics, caused by the discovery of “incommensurable line segments,” the Great mathematician Eu doxus had developed a theory of magnitudes, which was transformed later into mathematical measurement theory [3, 4], another fundamental theory of mathematics. This theory underlies all “continuous mathematics” includ ing differential and integral calculus. Influence of the “measurement problem” on the development of mathe matics is so great that the famous Bulgarian mathematician L. Iliev had de clared that “during the first epoch of mathematics development, from antiq uity to the discovery of differential and integral calculus, mathematics, inves tigating first of all the measurement problems, had created Euclidean geome try and number theory” [5]. Thus, the two “key” problems of ancient mathematics, the count problem and the measurement problem, had led to the formation of the two fundamen tal concepts of mathematics: natural numbers and irrational numbers—which, together with number theory and measurement theory, became the basis of “classical mathematics.” 4. Mathematics. The Loss of Certainty The book, Mathematics: The Loss of Certainty [6], written by American math ematician Morris Kline, had a huge influence upon the author and became a source of reflections about the nature and role of mathematics in modern sci ence; it is a pleasure for the author to retell briefly the basic ideas of Morris Kline’s book. Since the origin of mathematics as an independent branch of knowledge (Greek mathematics), and during more than two millennia, mathematics was engaged in a search for truth and had achieved outstanding successes. It seemed that the vast amount of theorems about numbers and geometrical figures, which was proved in mathematics, is an inexhaustible source of absolute knowledge which never can change. To obtain surprisingly powerful results, mathematicians had used a spe cial deductive method which allowed them to get new mathematical results (theorems) from a small number of axiomatic principles, named by axioms. The nature of the deductive method guarantees a validity of the conclusion if the initial axioms are true. Euclid’s Elements became the first great mathe matical work in this area, which is a brilliant example of the effective applica tion of the deductive method. Introduction xxiii Euclidean geometry became the most esteemed part of mathematics—not only because the deductive construction of mathematical disciplines had be gun with the Euclidean geometry—but, also its theorems completely corre sponded to the results of physical research. It was considered a firm scientific axiom for many millennia. Euclidean geometry is the geometry of the physical world surrounding us. That is why the unusual geometries created at the be ginning of the 19th century, named nonEuclidean geometries, became the first “blow” to the harmonious building of mathematical science. These un usual geometries had forced mathematicians to recognize that mathematical theories and theorems are not absolute truths in application to Nature. It was proved that new geometries are mathematically correct, that is, they could be geometrical models of the real world similar to Euclidean geometry, but then the following question arises: what geometry is a true model of the real world? Finding the contradictions in Cantor’s theory of infinite sets was another “blow” to mathematics. Comprehension of the “Tsarina of sciences” is not perfect regarding its structure; it lacks much, and it is subjected to monstrous contradic tions, which can appear at any moment; it shocked mathematicians. The reaction of mathematicians to all of these events was ambiguous. Unfortunately, the ma jority of mathematicians had simply decided to ignore these contradictions. In stead, they fenced themselves off from the external world and concentrated their efforts on the problems arising within the modern field of mathematics, that is, mathematicians decided to break connections with natural sciences. What was mathematics during several millennia? For previous genera tions, mathematics was first of all of the greatest creation of human intellect intended for nature’s research. The natural sciences were the flesh and blood of mathematics and it fed mathematicians with their vivifying juices. Mathe maticians willingly cooperated with physicists, astronomers, chemists, and en gineers in searching for the solution to various scientific and technical prob lems. Moreover, many great mathematicians of the past were often outstand ing physicists and astronomers. The mathematics was the “Tsarina” and si multaneously the “Servant” of natural sciences. Morris Kline noticed that “pure” mathematics, which had completely dis associated from the inquiries of natural sciences, was never the center of at tention and interest of the great mathematicians of the past. They considered “pure” mathematics as some kind of “entertainment,” a rest from much more important and fascinating problems, which were put forward by natural sci ences. In the 18th century, such abstract science like number theory had in volved only a few mathematicians. For example, Euler, whose scientific inter ests had been connected with number theory, was the first to be a recognized Alexey Stakhov MATHEMATICS OF HARMONY xxiv specialist in mathematical physics. The Great mathematician Gauss did not consider number theory as the major branch of mathematics. Many of his col leagues suggested that he solves The Great Fermat Theorem. In one letter, Gauss noted that Fermat’s hypothesis is an isolated mathematical problem, which is not connected with the most important mathematical problems, and conse quently, it is not of particular interest. Morris Kline specifies the various reasons that induced mathematicians to depart from studying the real world. Widening mathematical and natural scientific research did not allow scientists to feel equally free in both mathe matics and natural sciences. The problems, that stood before natural sciences a solution to which the great mathematicians of the past participated active ly nowadays became more and more complex, and many mathematicians had decided to limit their activity to the problems of “pure” mathematics. Abstraction, generalizations, specialization, and axiomatization are the basic directions of activity chosen by “pure” mathematicians. This activi ty led to the situation where, nowadays, mathematics and natural sciences go different ways. New mathematical concepts are developing without any attempt to find their applications. Moreover, mathematicians and repre sentatives of natural sciences do not understand each other today—owing to the excessive specialization in fields and often mathematicians do not understand each other. What can resolve this situation? Morris Kline emphasizes that researchers should return to nature and natural sciences, which were the original objectives of mathematics. Ultimately, common sense should win. The mathematical world should search for a distinction not between “pure” and applied mathematics, but between the mathematics; whereby, its purpose is to find a solution to rea sonable problems. Mathematicians should not indulge someone’s personal tastes and whims as our quests in mathematics is purposeful and neverending be cause mathematics is rich in content that is empty, alive, and bloodless. 5. A “Harmony Problem” As is known, returning to the past is a fruitful source of cognition to the present. The return to the sources of mathematics, to its history, is one of the impor tant directions to overcome the crisis of contemporary mathematics. In re turning to ancient science, particularly Greek science, we should pay atten tion to an important scientific problem, which was the focus of ancient sci ence starting with Pythagoras and Plato. Introduction xxv We are talking about the “harmony problem.” What is the harmony? Wellknown Russian philosopher Shestakov in his remarkable book Har mony as an Aesthetic Category [7] emphasizes that “in the history of aesthet ic doctrines, the diversified types of understanding of harmony were put forward. The concept of “harmony” is multiform and used extremely widely. It meant the natural organization of nature and space, a beauty of the hu man physical and moral world, principles of art works’ design or the law of aesthetic perception.” Among the various types of harmony (mathematic, aesthetic, artistic), which arose during the development of science and aes thetics, we will first be interested in mathematical harmony. In this sense, harmony is understood as equality or proportionality of the parts between themselves and the parts with the whole. In the Great Soviet Encyclopedia, we can find the following harmony definition, which expresses the mathe matical understanding of the harmony: “The harmony of an object is a pro portionality of the parts and the whole, a merge of the various components of the object to create a uniform organic whole. In harmony, the internal order and the measure of the object had obtained external revealing.” In the present book we concentrate our attention on mathematical harmo ny. It is clear that the mathematical understanding of harmony accepts, as a rule, the mathematical kind, and it is expressed in the form of certain numer ical proportions. Shestakov emphasizes [7] that mathematical harmony “fixes attention on its quantitative side and is indifferent to qualitative originality of the parts forming conformity... The mathematical understanding of the har mony fixes, first of all, quantitative definiteness of the harmony, but it does not express aesthetic quality of the harmony, its expressiveness, connection with a beauty.” 6. The Numerical Harmony of the Pythagoreans Pythagoreans, for the first time, put forth the idea of harmonious organiza tion of the universe. According to Pythagoreans, “harmony is an internal con nection of the things, without which the Cosmos could not exist.” At last, according to Pythagoras, harmony has numerical representation, namely that harmony is connected with the concept of number. The Pythagoreans had created the doctrine about the creative essence of number and their number theory had a qualitative character. Aristotle, in his “Metaphysics”, emphasiz es this feature of the Pythagorean doctrine: Alexey Stakhov MATHEMATICS OF HARMONY xxvi “The socalled Pythagoreans, studying mathematical sciences, for the first time have moved them forward and, basing on them, began to consider mathematics as the beginnings of all things... Because all things became like to numbers, and numbers occupied first place in all nature, they assumed that the elements of numbers are the begin ning of all things and that all universe is harmony and number.” Pythagoreans recognized that a form of the universe should be harmoni ous, and all elements of the universe are connected with harmonious figures. Pythagoras taught that the Cube originates the Earth, the Tetrahedron the Fire, the Octahedron the Air, the Icosahedron the Water, the Dodecahedron the sphere of the universe, that is, the Ether. The Pythagorean doctrine about the numerical harmony of the universe had influenced the development of all subsequent doctrines about the nature and essence of harmony. It was reflected upon and developed in the works of great thinkers. In particular, the Pythagorean doctrine underlies Plato’s cos mology. Plato developed the Pythagorean doctrine; specifically emphasizing the cosmic importance of harmony. He remained firmly convinced that world har mony can be expressed in numerical proportions. The influence of Pythagore ans is especially traced in Plato‘s “Timaeus”; whereby, Plato developed the doc trine about proportions and analyzed the role of Regular Polyhedrons (Platon ic Solids), from which—in his opinion—God had created the world. The main conclusion, which follows from the Pythagorean doctrine, con sists of the following. Numerical or mathematical harmony is objective prop erty of the universe, it exists irrespective of our consciousness and is expressed in the harmonious organization of all in the real world starting from cosmos and finishing by microcosm. 7. A “Harmony Problem” in Euclid’s Elements We ask how Pythagoras and Plato’s harmonious ideas were reflected in antique mathematics. To answer this question we analyze the greatest mathematical work of Greek mathematics: the Elements of Euclid. As is known, the Elements of Euclid is not an original work. A significant part of Elements was written by Pythagorean mathematicians. Their contribution to the theory of proportions— in which all ancient science and culture is based—is especially great. As the further progression of science had shown, the Pythagoreans, using numerical representations, did not leave the real world, but rather came nearer to it. Introduction xxvii The 13th, and final, Book of Euclid’s Elements is devoted to the theory of the regular polyhedrons, which is expressed in ancient science as universe har mony. The regular polyhedrons were used by Plato in his Cosmology and there fore they were named Platonic Solids. This fact originated the widespread hy pothesis formulated by Proclus—one of the most known commentators of Euclid’s Elements. According to Proclus’ opinion, Euclid created the Elements not with the purpose to present geometry as axiomatic mathematical science, but with the purpose to give the full systematized theory of Platonic Solids, in passing having covered some advanced achievements of the ancient mathe matics. Thus, the main goal of the Elements was a description of the theory of Platonic Solids described in the final book of Elements. It would not be out of place to remember that seemingly, the most important material of a scientific book is placed into the final Chapter of the book. Consequently, the place ment of the Platonic Solids theory in the final book of the Elements is indirect proof surrounding the validity of Proclus’ hypothesis; meaning that Pythago ras’ Doctrine about the numerical harmony of the universe got its brightest embodiment in the greatest mathematical work of the ancient science: Eu clid’s Elements. In order to develop a complete theory of the Platonic Solids, in particular the Dodecahedron, Euclid formulated in Book II the famous Theorem II,11 about the division in the extreme and mean ratio (DEMR), which is known in modern science under the name of the golden section. DEMR penetrated all Books of Euclid’s Elements, and it had been used by Euclid for the geometric construction of the following “harmonic” geometric figures: equilateral trian gle with the angles 72°, 72° and 36° (the “golden” equilateral triangle), regular pentagon and then the Dodecahedron based on the golden section. Taking into consideration Proclus’ hypothesis, and a role of the DEMR in Euclid’s Elements, we can put forward the following unusual hypothesis: Euclid’s Ele ments was the first attempt to create the “Mathematical Theory of Har mony” which was the main idea of Greek science. It is clear that the formulation of the division in the extreme and mean ratio (the golden section) can be considered as the “key” mathematical dis covery in the field of the “harmony problem.” The Great Russian philosopher Alexey Losev wrote in one of his articles that: “From Plato’s point of view, and generally from the point of view of all antique cosmology, the universe is a certain proportional whole that is subordinated to the law of harmonious division, the Golden Section.” Thus, we have to add the “harmony problem” to the list of the “key” prob lems of mathematics regarding the stage of its origin. Such approach leads us Alexey Stakhov MATHEMATICS OF HARMONY xxviii to the original view on the history of mathematics. This idea underlies the present book. During its historical development, the “classical mathematics” had lost Pythagoras’ and Plato’s “harmonious idea” embodied by Euclid in his Ele ments. As the outcome, mathematics had been divided into a number of math ematical theories (geometry, number theory, algebra, differential and integral calculus, etc.), which sometimes have very weak correlations. Unfortunately, a significance of the “golden mean” had been belittled in modern mathematics and theoretical physics. For many modern mathematicians, the “golden sec tion” reminds us of a “beautiful fairy tale,” which has no relation to serious mathematics. 8. Fibonacci Numbers Nevertheless, despite the negative relation of “materialistic” mathematics to the “golden mean,” its theory continued to develop. The famous Fibonac ci numbers: 1, 1, 2, 3, 5, 8, 13, 21, 34, …, had been introduced into mathe matics during the 13th century by the famous Italian mathematician Le onardo from Pisa (Fibonacci) at the solution of the rabbits reproduction problem. It is necessary to note that the method of recursive relations— one of the most powerful methods of combinatorial analysis—follows di rectly from Fibonacci’s discovery. Later, the Fibonacci numbers had been found in many natural objects and phenomena, in particular, the botanical phenomenon of phyllotaxis. 9. The First Book on the Golden Mean in the History of Science During the Italian Renaissance, interest in the “golden mean” arose with new force. Of course, the universal genius of the Italian Renaissance Leonardo da Vinci could not pass the division of the extreme and mean ratio (the golden section). There is an opinion that Leonardo had introduced into the Renais sance culture by the name of the “golden section.” Leonardo da Vinci had influenced the book Divina Proportione [8], which was published by Italian mathematician Luca Paccioli in 1509. This unique book was the first mathe matical book on the “golden mean” in history. The book was illustrated with 60 brilliant geometric figures drawn by Leonardo da Vinci; additionally, the book had a great influence on Renaissance culture. Introduction xxix 10. Johannes Kepler and the Golden Section In the 17th century, astronomer and mathematician Johannes Kepler created the original geometrical model of Solar system based on Platonic Solids. Ke pler had expressed his admiration of the golden section with the following words: “Geometry has two great treasures: one is the Theorem of Pythagoras; the other, the division of a line into extreme and mean ratio. The first, we may compare to a measure of gold; the second we may name a precious stone.” 11. Fibonacci Numbers and the Golden Section in 19th Century Science After Kepler’s death, interest in the golden section, considered one of the two “treasures of geometry,” decreased; whereby, such strange oblivion continued for two centuries. Active interest in the golden section revived in mathemat ics in the 19th century. During this period, many mathematical works were devoted to Fibonacci numbers and the golden mean, and according to the witty saying of one mathematician: they “started to reproduce as Fibonacci’s rabbits.” French mathematicians Lucas and Binet became the leaders of this type of research in 19th century. Lucas had introduced into mathematics the name “Fibonacci Numbers,” and also the famous Lucas numbers (1, 3, 4, 7, 11, 18, ...). Binet had deduced the famous Binet formulas, which connect the Fi bonacci and Lucas numbers with the golden mean. During this time, the German mathematician Felix Klein tried to unite together all branches of mathematics on the base of the Regular Icosahedron, the Platonic Solid—dual to the Dodecahedron. Klein treats the Regular Icosa hedron based on the golden section as the main geometric object, from which the branches of the five mathematical theories follow, namely, geometry, Ga lois’ theory, group theory, invariant theory, and differential equations. Klein’s main idea is extremely simple: “Each unique geometrical object is somehow or another connected to the properties of the Regular Icosahedron.” 12. The Golden Section and Fibonacci Numbers in Science of the 20th and 21st Centuries In the second half of the 20th century the interest in Fibonacci numbers and the golden mean in mathematics had revived with new force, and the revival expanded Alexey Stakhov MATHEMATICS OF HARMONY xxx into the 21st century with many original books [957] being published that were devoted to the golden mean, Fibonacci numbers, and other related topics, which is evidence of the increasing interest in the golden mean and Fibonacci numbers in modern science. Prominent mathematicians Gardner [12], Vorobyov [13], Coxeter [14], and Hoggatt [16] were the first researchers who felt new tenden cies growing in mathematics. In 1963, the group of American mathematicians had organized the Fibonacci Association and they started publishing the math ematical journal The Fibonacci Quarterly. Owing to the activity of the Fibonacci Association and the publications of the special books by Vorobyov [13], Hog gatt [16], Vaida [28], Dunlap [38], and other mathematicians, a new mathemat ical theory—the “Fibonacci numbers theory”—appeared in contemporary math ematics. This theory has its own interesting mathematical history, which is pre sented in the book A Mathematical History of the Golden Number, written by the prominent Canadian mathematician Roger HerzFishler [40]. In 1992 a group of the Slavic scientists from Russia, Ukraine, Belarus, and Poland had organized the socalled Slavic “Golden” Group. Resulting from the initiative of this group, the International symposiums of “The Golden Section and Problems of System Harmony” had been held in Kiev, Ukraine in 1992 and 1993, and then again in Stavropol, Russia from 19941996. The golden mean, pentagram, and Platonic Solids were widely used by astrology and other esoteric sciences, and this became one of the reasons for the negative reaction of “materialistic” science towards the golden mean and Platonic Solids. However, all attempts of “materialistic” science and mathe matics to forget and completely disregard the “golden mean” and Platonic Solids and to throw them out along with astrology and esoteric sciences on the “dump of the doubtful scientific concepts,” had failed. Mathematical models based on the golden mean, Fibonacci numbers, and Platonic Solids had proved to be “enduring,” and they began to appear unexpectedly in different areas of nature. Already, Johannes Kepler had found Fibonacci’s spirals on the surface of the phyllotaxis objects. The research of the phyllotaxis objects growth made by the Ukrainian architect Oleg Bodnar [37, 52] demonstrated that the ge ometry of phyllotaxis objects is based on a special class of hyperbolic func tions—the “golden” hyperbolic functions. In 1984, the Byelorussian, philoso pher Eduardo Soroko, had formulated the “Law of structural harmony of sys tems” [25]. This law confirmed a general character of selforganized processes in the system of any nature; it demonstrated that all selforganized systems are based on the generalized golden pproportions. Shechtman’s quasicrys tals, based on the Platonic icosahedron, and fullerenes (Nobel Prize of 1996), were based on the Archimedean truncated icosahedron, had confirmed Felix Introduction xxxi Klein’s great prediction about the fundamental role of the icosahedron in sci ence and mathematics [58]. Ultimately, Petoukhov’s “golden” genomatrices [59] did completed the list of modern outstanding discoveries based on the golden mean, Fibonacci numbers, and the regular polyhedra. It is possible to assume that the increasing interest in the golden mean and Fibonacci numbers in modern theoretical physics and computer science is one of the main features of 21st century science. Prominent theoretical phys icist and engineering scientist Mohammed S. El Nashie is a world leader in this field [6072]. El Nashie’s discovery of the golden mean in the famous phys ical twoslit experiment—which underlies quantum physics—became a source of many important discoveries in this area, in particular, the Einfinity theory. In this respect, we mention the works of El Nashie’s numerous followers work ing in theoretical physics [7383]. It is also necessary to note the contribution of Slavic researchers to this important area. The book [53] written by the Byelorussian physicist Vasyl Pertrunenko is devoted to the applications of the golden mean in quantum physics and astronomy. In 2006, the book Meta physics of the 21st century [57], edited by the famous Russian physicist and theorist Y.S. Vladimirov was published. The book [57] consists of three chap ters and the last chapter was devoted to the golden mean applications in mod ern science. This chapter begins with two important articles [59, 84]. Sta khov’s article [84] is devoted to the substantiation of “Harmony Mathemat ics” as a new interdisciplinary direction of modern science. Petoukhov’s arti cle [59] is devoted to the description of the important scientific discovery: the “golden” genomatrices; which reaffirms the deep mathematical connection between the golden mean and genetic code. The famous Russian physicist Professor Vladimirov (Moscow University) finishes his book Metaphysics [85] with the following words: “It is possible to assert that in the theory of elec troweak interactions there are relations that coincide with the ‘Golden Sec tion’ that play an important role in the various areas of science and art.” In the second half of the 20th century multiple interesting mathematical discoveries in the area of golden mean applications in computer science and mathematics had been made [86119]. In 1956, the young American mathe matician George Bergman made an important mathematical discovery in the field of number systems [86]. We are talking about the number system with irrational base (the golden mean) described in [86]. Modern mathematicians had been so anxious of overcoming the crisis in the basis of mathematics that they simply had not noticed Bergman’s discovery, which is, without doubt, one of the greatest mathematical discoveries in the field of number systems after the discovery by Babylonians of the positional principle of number repre Alexey Stakhov MATHEMATICS OF HARMONY xxxii sentation. Bergman’s number system was generalized by Alexey Stakhov who developed in the book [24] more general class of the number systems with irra tional radices named “Codes of the golden proportion.” Alexey Stakhov, in the article [105], developed a new approach to geometric definition of real numbers that is of great importance for number theory. In his article [87], and then in the book [20], Stakhov developed the socalled Fibonacci codes. The codes of the golden proportion and Fibonacci codes became a source of the Fibonacci com puter project [30] developed in the Soviet Union. This computer project was an original project, which was defended by 65 patents issued by the State Patent ing Departments of the United States, Japan, England, Germany, France, Can ada, and other countries [120131]. Parallel with Soviet computer science, work continued on Fibonacci computers in the United States [132135]. In the works [44, 103, 113, 114], a new class of square matrices, the generalized Fibonacci matrices and the socalled “golden” matrices, was developed. This led to a new kind of theory of coding and cryptography [44, 113, 114]. A new class of hyperbolic functions, the hyperbolic Fibonacci and Lucas functions, introduced by Alexey Stakhov, Ivan Tkachenko, and Boris Rozin [51, 98, 106, 116, 119], was another important modern mathematical discov ery. The beginning of the 21st century is characterized by a number of the interesting events; all of which have a direct relation to Fibonacci numbers and the golden mean. First of all, it is necessary to note that three Interna tional Conferences on Fibonacci Numbers and their Applications were held in the 21st century (Arizona, USA, 2002; Braunschweig, Germany, 2004; Cal ifornia, USA, 2006). In 2003, the international conference Problems of Har mony, Symmetry, and the Golden Section in Nature, Science and Art was held in Vinnitsa, Ukraine following the initiative of the Slavic “Golden” Group, which had transformed into the International Club of the Golden Section. In 2005, the Academy of Trinitarizm (Russia) and the International Club of the Golden Section, had organized the Institute of the Golden Section. Intersecting the 20th and 21st centuries, Western and Slavic scientists had published a number of scientific books in the field of the golden mean and its applications. The most interesting of them are the following: Dunlap R.A. The Golden Ratio and Fibonacci Numbers (1997) [38]. HerzFishler Roger. A Mathematical History of the Golden Number (1998) [40]. Vera W. de Spinadel. From the Golden Mean to Chaos (1998) [42]. Introduction xxxiii Gazale Midhat J. Gnomon. From Pharaohs to Fractals (1999) [45]. Kappraff Jay. Connections. The Geometric Bridge between Art and Sci ence (2001) [47]. Kappraff Jay. Beyond Measure. A Guided Tour Through Nature, Myth, and Number (2002) [50]. Shevelev J.S. Metalanguage of the Living Nature (2000) (Russian)[46]. Petrunenko V.V. The Golden Section in Quantum States and its Astro nomical and Physical Manifestations (2005) (Russian) [53]. Bodnar O.J. The Golden Section and NonEuclidean Geometry in Sci ence and Art (2005) (Ukrainian) [52]. Soroko E. M. The Golden Section, Processes of Selforganization and Evolution of System. Introduction into General Theory of System Har mony (2006) (Russian) [56]. Stakhov A.P., Sluchenkova A.A.. Scherbakov I.G. The da Vinci Code and Fibonacci Series (2006) (Russian) [55]. Olsen Scott. The Golden Section: Nature’s Greatest Secret (2006) [54]. This list confirms a great interest in the golden mean in 21st century science. 13. The Lecture: “The Golden Section and Modern Harmony Mathematics” By the end of the 20th century, the development of the “Fibonacci numbers theory” was widening intensively. Many new generalizations of Fibonacci numbers and the golden section had been developed [20]. Different unexpected applications of Fibonacci numbers and the golden section particularly in the oretical physics (the hyperbolic Fibonacci and Lucas functions [51, 98, 106]), computer science (Fibonacci codes and the codes of the golden proportion [20, 24, 87, 89, 102]), botany (the law of the spiral biosymmetries transforma tion [37]), and even philosophy (the law of structural harmony of systems [25]) were obtained. It became clear that the new results in this area were far beyond the traditional “Fibonacci numbers theory” [13, 16, 28]. Moreover, it became evident that the name “Fibonacci numbers theory” considerably nar rows the subject of this scientific direction—which studies mathematical mod els of system harmony. Therefore, the idea to unite the new results in the theory Alexey Stakhov MATHEMATICS OF HARMONY xxxiv of the golden mean and Fibonacci numbers and their applications under the flag of the new interdisciplinary direction of the modern science, named “Harmony Mathematics,” appeared. Such idea had been presented by Alexey Stakhov in the lecture “The Golden Section and Modern Harmony Mathematics” at the Seventh International Conference on Fibonacci Numbers and their Applications in Graz, Austria in July 1996. The lecture was later published in the book Appli cations of Fibonacci Numbers [100]. After 1996, the author continued to develop and deepen this idea [101 119]. However, the creation of “Harmony Mathematics” was a result of col lective creativity; the works of other prominent researchers in the field of the golden section and Fibonacci numbers Martin Gardner [12], Nikolay Voro byov [13], H. S. M. Coxeter [14], Verner Hoggat [16], George Polya [17], Alfred Renyi [23], Stephen Vaida [28], Eduardo Soroko [25, 56], Jan Grzedz ielski [26], Oleg Bodnar [37, 52], Nikolay Vasutinsky [31], Victor Korobko [43], Josef Shevelev [46], Sergey Petoukhov [59], Roger HerzFishler [40], Jay Kappraff [47, 50], Midhat Gazale [45], Vera W. de Spinadel [42], R.A. Dunlap [38], Scott Olsen [54], Mohammed S. El Nashie [6072], and other scientists had influenced the author’s research in this field. “Harmony Mathematics,” in its origin, goes back to the Euclidean problem of “division in the extreme and mean ratio” (the golden section) [40]. Harmony Mathematics is a continuation of the traditional “Fibonacci numbers theory” [13, 16, 28]. What are the purposes of this new mathematical theory? Similar to “classical mathematics,” which is defined sometimes as the “science about mod els” [5], we can consider Harmony Mathematics as the “science about the models of harmonic processes” in the world surrounding us. 14. Two Historical Ways of Mathematics Development In research, returning to the origin of mathematics, we can point out the two ways of mathematics development, which arose in the ancient mathematics. The first way was based on the “count problem” and the “measurement prob lem” [1]. In the period of mathematics origin two fundamental discoveries were made. The positional principle of number representation [2] was used in all known numeral systems, including the Babylonian sexagecimal, deci mal, and binary systems. Ultimately, the development of this direction culmi nated in the formation of the concept of natural numbers; it also led to the creation of number theory—the first fundamental theory of mathematics. In commensurable line segments discovered by Pythagoreans led to the dis Introduction xxxv covery of irrational numbers and the creation of measurement theory [3, 4]—which was the second fundamental theory of mathematics. Ultimate ly, natural and irrational numbers became those basic mathematical concepts that underlie all mathematical theories of “classical mathematics,” including num ber theory, algebra, geometry, and differential and integral calculus. Theoretical physics and computer science are the most important applications of “classical mathematics” (see Fig. I.1). The “key” problems of the ancient mathematics Count problem Measurement problem Harmony problem Positional principle of number representation Incommensurable line segments Division in the extreme and mean ratio Number theory and natural numbers Measurement theory and irrational numbers Theory of Fibonacci numbers and the golden section Classical mathematics Theoretical physics Computer science Harmony Mathematics The “golden” theoretical physics The “golden” computer science Figure I.1. Three “key” problems of the ancient mathematics and new directions in mathematics, theoretical physics and computer science However, parallel with the “classical mathematics” in the ancient science Alexey Stakhov MATHEMATICS OF HARMONY xxxvi another mathematical theory—Harmony Mathematics—had started to devel op. Harmony Mathematics originated from another “key” idea of antique sci ence—the “Harmony problem.” The Harmony Problem triggered the “Doc trine about Numerical Harmony of the Universe” that was developed by Pythagoras. A division in the extreme and mean ratio (the golden section) was the “key” mathematical discovery in this area [40]. The development of this idea resulted in the Fibonacci numbers theory [13, 16, 28] in modern mathemat ics. However, the extension of the Fibonacci numbers theory and its applica tions coupled with the generalization of Fibonacci numbers and the golden section produced the concept of “Harmony Mathematics” [100] as a new interdisciplinary direction of modern science and mathematics. This can re sult in the creation of the “golden” theoretical physics, based on the “gold en” hyperbolic models of nature [51, 98, 106, 116, 118], and the “golden” computer science, based on new computer arithmetic [20, 24, 30, 87, 89, 94, 104] and a new theory of coding and cryptography [55, 113, 114]. 15. The Main Goal of the Present Book It seems that the dramatic history of the DEMR (the golden section) that continued over several millennia has ended as a great triumph for the golden section in the beginning of the 21st century. Many outstanding scientific dis coveries that are based on the golden section (quasicrystals, fullerenes , “gold en” genomatrices and so on) gave reason to conclude that the golden section may be considered as some kind of “metaphysical knowledge,” “prenum ber,” or “universal code of Nature,” which could become the basis for the future development of science; particularly, theoretical physics, genetics, and computer science. This idea is the main concept of the book [57] and the articles [59, 84]. These scientific facts demand reappraisal of the role of the golden section in contemporary mathematics. The main purpose of the present book is to revive the interest in the gold en section and Pythagoras, Plato, and Euclid’s “harmonic idea” in modern mathematics, theoretical physics, and computer science. It also strives to dem onstrate that the Euclidean problem of the “division in extreme and mean ratio” (the golden section) is a powerful and fruitful source of many funda mental ideas and concepts of contemporary mathematics, theoretical physics, and computer science. We consider different generalizations of the golden mean, in particular, the generalized golden pproportions (p=0, 1, 2, 3, …) and Introduction xxxvii the generalized golden means of the order m (m is a positive real number) as fundamental mathematical constants similar to the numbers π and e. We show that this approach resulted in: a new class of elementary functions—the hy perbolic Fibonacci and Lucas functions; a new class of the recursive numerical sequences—the generalized Fibonacci and Lucas pnumbers (p=0, 1, 2, 3, …) and the generalized Fibonacci and Lucas numbers of the order m (m is a positive real number); and it also led to a new class of square matrices—the Fibonacci and “golden” matrices. Also, this approach resulted in a new measurement the ory, algorithmic measurement theory, in a new class of number systems with irrational radices that are codes of the golden pproportions. Additionally, a new kind of computer arithmetic, the Fibonacci and “golden” arithmetic and the ternary mirrorsymmetric arithmetic, was developed, as well as a new cod ing theory based on the Fibonacci matrices and a new kind of cryptography— the “golden” cryptography. The book consists of three parts. Part I “Classical Golden Mean, Fibonacci numbers, and Platonic Solids” consists of three chapters, Chapter 1 “The Gold en Section”, Chapter 2 “Fibonacci and Lucas Numbers”, and Chapter 3 “Regu lar Polyhedrons”. Part I is popular introduction into the Fibonacci numbers theory and its applications. Part I is intended for a wide audience including mathematics teachers of secondary schools, students of colleges and universi ties. Also, Part I can attract attention to the representatives of various branches of modern science and art that are interested in both creative and practical ap plications of the golden mean, Fibonacci numbers, and Platonic Solids. Part II “Mathematics of Harmony” consists of three chapters, Chapter 4 “Generalizations of Fibonacci Numbers and the Golden Mean,” Chapter 5 “Hyperbolic Fibonacci and Lucas Functions,” and Chapter 6 “Fibonacci and Golden Matrices”. Part II calls for special knowledge in mathematics and is intended, first of all, for mathematicians and scientists in theoretical physics. Part III “Applications in Computer Science” consists of five chapters, Chapter 7 “Algorithmic Measurement Theory”, Chapter 8 “Fibonacci Com puters”, Chapter 9 “Codes of the Golden Proportion” , Chapter 10 “Ternary MirrorSymmetrical Arithmetic,” and Chapter 11 “A New Coding Theory Based on Matrix Approach.” Part III is intended for mathematicians and spe cialists in computer science. Note that Parts II and III are, in the main, results of original researches obtained by the author in about 40 years of scientific work. This page intentionally left blank Acknowledgements xxxix Acknowledgements This book is a result of my lifetime research in the field of the Golden Section, Fibonacci numbers, “Mathematics of Harmony,” and their applications in mod ern science. I have met along the way a lot of extraordinary people who helped me to broaden and deepen my scientific knowledge and professional skills. More than 40 years ago I read the remarkable brochure Fibonacci Numbers written by the Russian mathematician Nikolay Vorobyov. This brochure was the first mathematical work on Fibonacci numbers published in the second half of the 20th century. It determined my scientific interests in Fibonacci numbers. In 1974, I met with Professor Vorobyov in St. Petersburg (formerly Leningrad) and narrated him about my scientific achievements in this area. He gave me as a keepsake his brochure Fibonacci Numbers with the following inscription: “To highly respected Alexey Stakhov with Fibonacci’s greetings.” I have great gratitude to my teacher, the outstanding Ukrainian scientist, Professor Alexander Volkov (19242007), under whose leadership I defend ed my PhD dissertation in 1966 and DrSci dissertation in 1972. These disser tations were the first step in my research, which led me to the new scientific direction Mathematics of Harmony. I appreciated a true value of mathematics as a “method of thinking” after an influential collaboration with Igor Vitenko, a graduate of the mathemati cal faculty of the Lvov University, who, without any doubts, would have been known as an outstanding mathematician and glory of Ukrainian science if he hadn’t passed away so early in life. His death in October 1974—which hap pened as a result of suicide—was a big loss. Also, I consider the outstanding Russian philosopher George Chefranov as my teacher of philosophy. Our evening walks, along Chekhov’s street of Taganrog city, were accompanied by hot discussions of philosophical problems of science, which became a major philosophical university for me. In my stormy scientific life I met many fine people, who could understand me, my enthusiasm, and appreciate my scientific direction. With deep grati tude, I recall a meeting in the Austrian city of Graz in 1976 with the excep tional Austrian mathematician Alexander Aigner. His review of my lecture at Graz University was the beginning of international recognition of my sci entific direction. The German scientist and academician Dr. Volker Kempe who now lives in Austria, also impacted on my career. Thanks to Dr. Kempe, I was honored to be invited to Dresden Technical University as a visiting Alexey Stakhov MATHEMATICS OF HARMONY xl professor in 1988. Another remarkable person who had a great influence on my work was the Ukrainian mathematician and academician Yury Mitropol sky (19172008). His influence on my research pertinent to the history of mathematics and other topics, such as the application of “Harmony Mathe matics” to the contemporary mathematics and to the mathematical educa tion, is inestimable contribution. Thanks to the support of Yury Mitropolski, I was able to publish many scientific articles in various Ukrainian academic journals. My arrival in Canada in 2004 set the arena for the next stages of develop ment in my scientific researches. Thanks to support from the famous physicist Professor Mohammed El Nashie, the editorinchief of the international mag azine Chaos, Solitons and Fractals, I was able to publish many articles in this remarkable interdisciplinary journal. Within two years, I published 14 funda mental articles in Chaos, Solitons and Fractals, which closed my cycle of re search in the field of “Mathematics of Harmony.” These articles attracted at tention of the international scientific community to my latest scientific re sults; also, these articles paved the way for making new professional contacts with worldrenowned Western scientists. Scientific discussion with outstand ing scientists in the field of the Golden Section, such as the Argentinean math ematician Vera W. de Spinadel—author of the book From the Golden Mean to Chaos (1998), the American mathematician Louis Kauffman—editor of the Knots and Everything Series from World Scientific, the American researcher Jay Kappraff—author of two remarkable books Connections: The geometric bridge between Art and Science (2001) and Beyond Measure: A Guided Tour Through Nature, Myth, and Number (2002), have considerably influenced on my research in the field of the Golden Section and Mathematics of Harmony. Thanks to the support from Professors Louis Kauffman and Jay Kappraff, this book was accepted by World Scientific for publication. I would like to express my gratitude to the initiative of Professor Vera W. de Spinadel, who helped with acceptance of my lecture, the “Three ‘Key’ Problems of Mathe matics on the Stage of its Origin and the “Harmony Mathematics” as Alterna tive Way of Mathematics Development,” as the plenary lecture at the Fifth Mathematics & Design International Conference (July 2007, Brazil). Furthermore, I would like to express special thanks to Dr. Scott Olsen, the American philosopher and prominent researcher in the golden mean and author of the book The Golden Section: Nature’s Greatest Secret (2006), for his invaluable assistance in the scientific and English editing of my book. Also, I would like to express my greatest gratitude to the Russian businessman and scientist Dr. Ivan Raylyan, a great enthusiast of the Golden Section, for his Acknowledgements xli financial assistance, which provided me with all of the best technology, infor mational tools, and pecuniary means to produce a cameraready manuscript. During my Canadian period of my scientific life I published a number of important articles in the field of the Harmony Mathematics together with Boris Rozin. I would like to express to him great gratitude for scientific col laboration. I would like to express my special thanks to Alexander Soypher for his invaluable assistance in overall design of the book and illustrations, cover de sign and color appendix “Museum of Harmony and Golden Section.“ Lastly, this book would never have been written without selfdenying sup port of my wife Antonina, who always creates perfect conditions for my sci entific work in any countries I have been; and who has been sailing with me for more than 47 years on my “Golden” journey. In addition, I would like to express my special appreciation to my daughter Anna Sluchenkova for her critical remarks, and her invaluable help in the English translation and edit ing of the book, and, especially, for her work in preparing illustrations, and coordination and final preparation of cameraready manuscript. Without her support this book would never been published. Alexey Stakhov This page intentionally left blank Credits and Sources xliii Credits and Sources Cover Cover design credit: Alexander Soypher Jacopo de Barbari, Portrait of Luca Pacioli source: Wikimedia Commons, Galaxy M74 see credit for Figure 2.6b, Nautilus spiral see credit for Figure 2.6a Ch.1. p.15 Niels Abel portrait and A monument of Niels Abel in Oslo source: www.abelprisen.no Figure 1.6 Credit card image source: www.automotoportal.com Figure 1.16 The Pentagon source: www.directionsmag.com Figure 1.19 Pentagonal symmetry in nature credit Alexander Soypher Figure 1.20 Hesire Panel from Shevelev I., Marutaev M., Shmelev I., Golden Section, 1990 Figure 1.21 Cheops Pyramid in Giza source: www.bibleplaces.com/giza.htm Figure 1.23 Doryphoros by Polycletes source: www.goldensection.eu/abb10.jpg Figure 1.24 Venus de Milo source: www.theoi.com Figure 1.25 Holy Family by Michelangelo source: britton.disted.camosun.bc.ca/goldslide/gold22.jpg Figure 1.26 Crucifixion by Rafael Santi source: britton.disted.camosun.bc.ca/goldslide/gold24.jpg Ch.1. p.44 Leonardo da Vinci from Swedish journal Svenska FamiljJournalen (18641887). Ch.1. p.51 Memorial stone in Borgo San Sepolcro and In honor of 500 anniversary of Summa composite photo credit Alexander Soypher; pictures sources: utenti.quipo.it/base5/pacioli/statuapacioli.htm, jeff560.tripod.com/stamps.html, numismaticaitaliana.lamoneta.it/moneta/WITL/35 Figure 1.30 Saint Pietro in Montorio source: www.wga.hu/art/b/bramante/ztempiet.jpg Alexey Stakhov MATHEMATICS OF HARMONY xliv Figure 1.31 Smolny Cathedral St.Petersburg, Russia source: www.aditours.com/russiafinland.htm Figure 1.33 Le Corbusier’s Modulor source: britton.disted.camosun.bc.ca/goldslide/gold39.jpg Figure 1.34 The Building of the United Nations Headquarters in New York source: www.inetours.com/New_York/Images/UN/UN_8874.jpg Figure 1.35 Near the window by Konstantin Vasiliev source: varvar.ru/arhiv/gallery/russian/vasilyev Figure 1.36 The golden mean in human body (a) from Kovalev F., Golden Section in Painting The golden mean in clothes design (b) source: www.olish.ru/picture/gold7.jpg Figure 1.37 Harmonic analysis of woman face source: www.beautyanalysis.com Figure 1.38 Harmonic analysis of Nefertiti’s portrait from Soroko E., Golden code in sculpture portrait of Nefertiti www.trinitas.ru/rus/doc/0232/004a/02320038.htm Figure 2.1 Fibonacci’s rabbits source: www.geider.net/esp/varios/numero_de_oro.htm Figure 2.2 Family tree of honeybees source: www.obscurious.co.uk/html/html.pages/Fibonacci.html Figure 2.6a Nautilus spiral this Wikipedia and Wikimedia Commons image is from the user Chris 73 and is freely available at commons.wikimedia.org/wiki/File: NautilusCutawayLogarithmicSpiral.jpg under the creative commons ccbysa 2.5 license. Figure 2.6b Galaxy M74 photo credit: Gemini Observatory/AURA, source: www.gemini.edu Figure 2.6c Sea Shell source: www.femorale.com.br Figure 2.8 Crystal’s spatial lattices of NaCl (а) and CaO (b) source: www.3dchem.com Figure 2.11 Helical symmetry source: Dr. Ron Knott’s web page www.mcs.surrey.ac.uk/Personal/R.Knott/Fibonacci/fibnat.html Figure 2.13 Phyllotaxis structures: Credits and Sources xlv (a) Cactus source: flickr.com/photos/wurzle/3668992 (b) Sunflower photograph by Janet Lanier from www.athousandandone.com/photos/2/44b99d279626a_s.jpg (c) Coneflower source: www.wunderground.com/data/wximagenew/m/mentalno mad/26.jpg (d) Romanescu source: morgana.djo.tudelft.nl/romanescu.jpg (e) Pineapple permission to use is granted by the GNU Free Documentation License (f) Pine cone source: flickr.com/photos/mrpattersonsir/11199034 Figure 2.14 Geometric models of phyllotaxis structures source: Jill Britton’s web page britton.disted.camosun.bc.ca/fibslide/jbfibslide.htm Ch.2. pp.130132 Edouard Zeckendorf, Verner Emil Hoggatt, Alfred Brousseau, Herta Taussig Freitag source: www.mscs.dal.ca/Fibonacci/biographies.html Figure 3.4 Platonic Solids source: Wikimedia Commons, permission to use is granted by the GNU Free Documentation License Figure 3.5 Archimedean solids source: Wikimedia Commons, permission to use is granted by the GNU Free Documentation License Figure 3.6 A construction of the Archimedean truncated icosahedron from E.Katz, Art and science – about polyhedrons in general and about truncated icosahedron in particular Figure 3.7 Cuboctahedron and icosidodecahedron source: Wikimedia Commons, permission to use is granted by the GNU Free Documentation License Figure 3.8 Rhombicuboctahedron and rhombicosidodecahedron source: Wikimedia Commons, permission to use is granted by the GNU Free Documentation License Figure 3.9 Snub hexahedron and snub dodecahedron source: Wikimedia Commons, permission to use is granted by the GNU Free Documentation License Figure 3.10 KeplerPoinsot solids source: Wikimedia Commons, this image is licensed under the Creative Commons Attribution ShareAlike 2.5 License. Figure 3.15 Egg of Life source: www.floweroflife.org/toverview6.htm Ch.3. p. 169 Dan Shechtman source: pard.technion.ac.il Alexey Stakhov MATHEMATICS OF HARMONY xlvi Figure 3.21 Buckminster Fuller source: users.midwestmail.com/nightlife/ cdalerocks/nightlife/zzzPermanent/030821BuckyFuller.html Fuller’s geodesic dome source: www.jcsolarhomes.com/image/dome1.gif Figure 3.22 The Montreal Biosphere source: www.ec.gc.ca/EnviroZine/images/Issue59/Biosphere_l.jpg Figure 3.23 The U.S. postage stamp source: www.bfi.org Figure 3.25 Molecule C60 source: www.3dchem.com/imagesofmolecules/c60.jpg Figure 3.31 Piero della Francesco’s “The Baptism of Christ” the compilation copyright is held by The Yorck Project and licensed under the GNU Free Documentation License Figure 3.32 Fra Giovanni da Verona’s intarsia source: www.flickr.com/photos/powersmitchell/2249856120 www.flickr.com/photos/powersmitchell/2249856176 Figure 3.34 Abstract art of Astrid Fitzgerald permission to use is granted by Astrid Fitzgerald, source: www.astridfitzgerald.com Figure 3.35 The art works of Teja Krasek permission to use is granted by Teja Krasek, source: tejakrasek.tripod.com Figure 3.36 The Geometric Art of John Michell source: www.thehope.org/jm_art.htm (Every attempt was made to contact the author for permission, Mr. Michell died on April 24, 2009 in Poole, England) Figure 3.37 Quantum Connections by Marion Drennen permission to use is granted by Marion Drennen, source: quantumconnectionsart.blogspot.com Ch.6. p.318 James Joseph Sylvester source: wwwhistory.mcs.stand.ac.uk/history/PictDisplay/Sylvester.html Ch.8. p.422 John von Neumann credit United States Department of Ener gy, source: Wikimedia Commons Ch.11. p.570 Claude Elwood Shannon credit Bell Labs, source: www.belllabs.com/news/2001/february/26/1.html Ch.11. p.573 Richard Wesley Hamming credit Samuel P. Morgan, source: cm.belllabs.com/cm/cs/alumni/hamming/index.html Credits and Sources xlvii Appendix Museum of Harmony and Golden Section Symmetry in Nature Butterfly credit Stavros Markopoulos, source: www.flickr.com/photos/markop/273706600 Fern leaf credit Ken Ilio, source: www.flickr.com/photos/ kenilio/217816458 Tiger credit Brian Mckay, source: www.flickr.com/photos/ 30775272@N05/2884963755 Peacock credit Gary Burke, source: www.flickr.com/photos/ gburke6/1806285477 Beetle credit Ivan Zheleznov, source: www.naturelight.ru/ show_photo/author429/9903.html Cow credit Steve Arens, source: www.flickr.com/photos/ stevenarens/410707104 Black Swan credit Francesca Birini, source: www.flickr.com/photos/firenzesca/1429628428 Kitten credit Erkhemchukhal Dorj, source: www.flickr.com/ photos/mongol/311285917 Cobra Snake credit Michael Allen Smith, source: www.flickr.com/photos/digitalcolony/2289243736 Pentagonal symmetry in nature Sea Urchin (on the background) underwater photo credit Cornelia Kraus Urchin shell source: www.fontplay.com/freephotos/imagesn/ fpfreefoto1519.jpg Passion flower source: www.passionflowers.co.uk Comb Sand Star credit Peter Bryant, source: nathistoc.bio.uci.edu/Echinos/Astropecten.htm Stapelia source: www.flickr.com/photos/21288646@N00/ 285588504 Sea Star credit Paul Shaffner, source: www.flickr.com/photos/ paulshaffner/2232846436 Red Frangipani source: www.fontplay.com/freephotos/ 4thfoldern/fp11040520.jpg Alexey Stakhov MATHEMATICS OF HARMONY xlviii Crown flower credit Jessy C.E, source: www.flickr.com/ photos/tropicaliving/1515336212 Periwinkle (Vinca minor) source: www.fontplay.com/ freephotos/nine/fpx04210712.jpg White Frangipani source: www.fontplay.com/freephotos/ 4thfoldern/fp11050518.jpg Hoya Carnosa source: http://www.flickr.com/photos/flatpix/ 1258040465 Golden spirals in nature Barred Spiral Galaxy NGC 1300 credit NASA, ESA, and The Hubble Heritage Team (STScI/AURA), source: hubblesite.org Sea shell source: www.flickr.com/photos/fdecomite/ 2985635111 Fiddlehead fern credit James W. Gentry Jr., source: www.flickr.com/photos/saynine/2348635947 Fern spiral (red) credit Faith Wilso, source: www.flickr.com/ photos/terrabella/2713519265 Nautilus shell see credit for Figure 2.6a Aloe spiral credit J. Brew, source: www.flickr.com/photos/ brewbooks/184343329 Fiddleheads (green) credit Dave & Lesley, source: www.flickr.com/photos/fireflies604/2834432521 Cala lily credit Dawn Endico, source: www.flickr.com/ photos/candiedwomanire/2886355 Snail source: www.flickr.com/photos/sailorganymede/ 2884930696 Rose credit Phil Price, source: www.flickr.com/photos/ philprice/2610013436 Whirlpool Galaxy M51 credit NASA, ESA, S. Beckwith (STS cI), and The Hubble Heritage Team (STScI/AURA), source: hubblesite.org The Spiralof Life credit Mr. Theklan, source: www.flickr.com/ photos/theklan/1161283822 Credits and Sources xlix Fibonacci numbers and phyllotaxis Data palm crown credit Steve Flippen, source: ww.flickr.com/ photos/sdflip/307295569 Aloe flower credit Dan Martin, source: www.flickr.com/ photos/acurazinedan/386035959 Monstera Deliciosa flower credit Luidgi, source: www.flickr.com/photos/luigistrano/2063302304 Pineapple source: puregreeni.us/wpcontent/uploads/2008/ 02/450pxpineapple1.jpg Banksia coccinea source: blog.paran.com/swan555/21036540 Norfolk Island Pine credit Brenda, source: www.flickr.com/ photos/omnia/67857983 Salak source: commons.wikimedia.org/wiki/File:Salak_01.jpg Cycad cone credit Glenn Fleishman, source: www.flickr.com/ photos/glennf/20342176 Golden Section in architecture United Nations Building source: www.inetours.com/ New_York/Pages/photos/UNHeadquartersbuilding.html The Pentagon, US Department of Defense building credit US Department of Defense, source: www.defenselink.mil/photos/ newsphoto.aspx?newsphotoid=1309 Smolny Cathedral in St. Petersburg source: www.lode.by/i/ photo/the_smolny_cathedral_174864.jpg St. Andrew Cathedral in Kiev source: www.tour2kiev.com/ pub_img/0704161414_st_andrew.jpg San Pietro in Montorio source: Wikimedia Commons en.wikipedia.org/wiki/San_Pietro_in_Montorio Notre Dame in Paris credit Andrea Schaffer, source: www.flickr.com/photos/aschaf/2836231278 Cathedral of Our Lady of Chartres credit Dimitry B, source: www.flickr.com/photos/ru_boff/520034494 Parthenon source: leftistmoon.files.wordpress.com/2007/10/ parthenon.jpg Great pyramids of Giza credit Ricardo Liberato, this image is Alexey Stakhov MATHEMATICS OF HARMONY l licensed under Creative Commons Attribution ShareAlike 2.0 License, source: en.wikipedia.org/wiki/File: All_Gizah_Pyramids.jpg The Great Pyramid of Khufu source: whgbetc.com/mind/ pyramids_giza.jpg Golden Section in Contemporary Abstract and Applied Art Polyhedral Universe credit Teja Krasek No. 258 from Cosmic Measures collection credit Astrid Fitzgerald Nine credit Marion Drennen Double Pentagram Pentagons credit John Michell Cosmic Gems credit Teja Krasek Pentagonal Hexagons credit John Michell No. 134 from Constructions collection credit Astrid Fitzgerald Life, the Universe, and Everything credit Teja Krasek No. 283 from All is Number: Number is All collection credit Astrid Fitzgerald One credit Marion Drennen Four credit Marion Drennen No.129 from Constructions collection credit Astrid Fitzgerald Homage to Pythagoras credit Marion Drennen This page intentionally left blank Part I Part I. Classical Golden Mean, Fibonacci Numbers, and Platonic Solids Alexey Stakhov THE MATHEMATICS OF HARMONY 2 Chapter 1 The Golden Section 1.1. Geometric Definition of the Golden Section 1.1.1. A Problem of the Division in the Extreme and Mean Ratio (DEMR) The Elements of Euclid is one of the best known mathematical works of ancient science. Written by Euclid in the 3rd century B.C., it contains the main theories of ancient mathematics: elementary geometry, number theory, algebra, the theory of proportions and ratios, methods of calculations of areas and volumes, etc. Euclid, in this work, systematized a 300year period of de velopment of Greek mathematics, and this work created a strong base for the further development of mathematics. The information about Euclid himself is extremely scanty. Except for several jokes, we only know that he taught at the mathematical school in Alexandria. The Elements of Euclid surpassed all works of his predecessors in the field of geometry, and during more than two millen nia, The Elements remained the basic work for the teaching of “Elementary Mathematics.” The 13 books of The Elements are dedicated to the knowledge of geometry and arithmetic of the Euclidean epoch. From The Elements of Euclid, the following geometrical problem, which was named the problem of “Division in Extreme and Mean Ratio” (DEMR), came to us [40]. This problem was formulated in Book II of The Elements as follows: Theorem II.11 (the area formulation of DEMR). To divide a line AB into two segments, a larger one AC and a smaller one CB so that: S ( AC ) = R ( AB,CB ) . (1.1) Note that S ( AC ) is the area of a square with a side AC and R ( AB,CB ) is the area of a rectangle with sides AB and CB. Chapter 1 The Golden Section 3 We can rewrite the expression (1.1) in the following form: S(AC) ( AC )2 = AB ×CB. (1.2) R(AB, BC) Now, divide both parts of the expression (1.2) by AC and then by CB. The expression (1.2) then takes the form of the following proportion: AB AC = . (1.3) A C B AC CB This form is wellknown to Figure 1.1. A geometrical interpretation of Theorem II.11 (The Elements of Euclid) mathematicians as the “golden section.” We can interpret a proportion (1.3) geometrically (Fig. 1.2): divide a line AB at the point C into two segments, a larger one AC and a smaller one CB, so that the ratio of the larger segment AC to the smaller segment CB is equal to the ratio of the line AB to the larger segment AC. A C B Figure 1.2. The division of a line in extreme and mean ratio (the golden section) Denote a proportion (1.3) by x. Then, taking into consideration that AB = AC + CB , the proportion (1.3) can be written in the following form: AB AC + CB CB 1 1 = = 1+ = 1+ = 1+ , AC x AC AC AC CB whence we obtain the following algebraic equation: x= (1.4) x 2 = x + 1. It follows from the “geometrical meaning” of the proportion (1.3) that the required solution of Eq. (1.4) has to be a positive number; it also follows that a positive root of Eq. (1.4) is a solution of the problem. If we denote this root by τ, then we obtain: 1+ 5 (1.5) . 2 This number is called the Golden Proportion, Golden Mean, Golden Num ber, or Golden Ratio. τ= Alexey Stakhov THE MATHEMATICS OF HARMONY 4 We can write the following identity, which connects the powers of the golden ratio: (1.6) τn = τn −1 + τn −2 = τ × τn −1 . The approximate value of the golden proportion is: τ ≈ 1.618033988749894848204586834365638117720309180…. Do not be astonished by this number! Do not forget that this number is an irrational one! In this book, we will use the following approximate value: τ ≈ 1.618 or even τ ≈ 1.62 . This surprising number, which possesses unique algebraic and geometrical properties, became an aesthetic canon of ancient Greek art and Renaissance art. Why did Euclid formulate Theorem II.11? As is shown in [40], by using this theorem, Euclid introduced the geometric construction of the golden triangle, pentagram, and dodecahedron (these geometric figures will be discussed below). 1.1.2. The Origin of the Concept and Title of the Golden Section Authorities vary over who introduced both the concept and terminology for the golden section. According to [40], the concept of the DEMR (or the golden section) was introduced in The Elements (see Theorem II.11). However, The Ele ments is not a completely original work. There is an opinion that the majority of theorems presented in The Elements are scientific results obtained by the Pythagore ans. Roger HerzFischler wrote [40]: “Many authors point to references according to which the pentagram was the symbol of the Pythagoreans, and from this, they deduce that the Pythagoreans were probably acquainted with DEMR.” Herz Fischler notes that the famous historians of mathematics, Heath and Van der Waer den, supported this point of view. However, Euclid did not use the term “golden section” in his works; instead, he used the term DEMR. In [40], we can trace the history of terminology for the DEMR throughout the ages. As follows from [40], the terminology of DEMR was used by Euclid, Fibonacci (1220), Zumberti (1516), Gryaneus (1533), Candalla (1566), Billingsley (1570), Commandino (1572), Clavius (1579), and Barrow (1722). However, many mathematicians used the term “Mid dle and Two Ends,” in particular: Abu Kamil (850930), alBiruni (9731050), Ger hard of Cremona (12th century), Adelard (12th century), Campanus of Novara (13th century), and Billingsley (1570). In addition to the definitions of the “divi sion in the extreme and mean ratio” and “proportion having a middle and two ends,” other definitions, namely: “divina proportione,” “proportionally divided,” “contin uous proportion,” “medial section,” “the golden number,” and “the golden section” were used. The term “divina proportione” was the title of Pacioli’s book published in 1509. This term was used by Kepler in his 1608 letter. We can find the term Chapter 1 The Golden Section 5 “proportionally divided” in Clavius’ 1574 edition of The Elements. Kepler, in his Mysterium Cosmographycum used the term “proportional division” for the DEMR. The term “continuous proportion” was used by Euclid and we can also find it used in Kepler’s 1597 letter. The term “medial section” was used by Leslie (1820) in connection with the DEMR. The “golden number,” the “golden section,” or the “golden mean” are the most popular names for the DEMR. As is emphasized in [40], Tannery (1882) used the term “Section d’Or” and Cantor (1894) used a title of the “golden shnitt.” Also noted in [40], for the first time, the term “golden schnitt” (the “golden section”) was used by Ohm in the second edition of his book Pure Elementary Mathematics (1835). However, other points of view exist about the origin of the term “the golden section” — mainly in Russian literature. Edward Soroko, who is one of the most authoritative Slavic researchers of the “golden section,” wrote in his book [25]: “The title of the “Golden Section” (“Sectio Aurea”) takes its origin from Claudia Ptolemey who gave this title for the number 0.618, after he had been convinced that the growth of a personwith perfect constitution is divided naturally in this ratio. The given title was fixed and then became popular due to Leonardo da Vinci who often used this title.” Unfortunately, the claim about the involvement of Leonardo da Vinci to the introduction of the title of the golden section has been questioned be cause there are no direct references to this title in his works. However, we should not forget that the great Italian mathematician Luca Pacioli was first to introduce the title “divine proportion” for the DEMR. Leonardo da Vinci actively participated in Pacioli’s book De Divina Proportione as he illustrated this unique mathematical work — which was the first book on the golden mean in world history. This means that, without any doubt, Leonardo da Vinci knew the concept of the “divine proportion.” Very often, the golden mean is denoted by the Greek letter F (the number PHI). This letter is the first one in the name of the wellknown Greek sculp tor Phidias, who widely used the golden section in his sculptural works. Re member that Phidias (5th century B.C.), together with Polyclitus, were con sidered to be two of the most famous and authoritative masters of ancient Greek sculpture of the Classical epoch. He became famous thanks to his su pervision over designing the Acropolis. He created the enormous bronze statue of Athena (“Winner in Battle”) in commemoration of the victory over the Persians. Also, he created two grandiose statues from gold and ivory: Athena Parthenos (“Maidens”) for the Parthenon and Zeus for the Olympian temple of Zeus (about 430 B.C.), which is considered one of the “Seven Miracles of Alexey Stakhov THE MATHEMATICS OF HARMONY 6 the World.” Despite the monumental character of his sculptures — which were unprecedented by their sizes among all Greek sculptures of that time (for example, the 9meter sculpture of Athena Parthenos or the 13meter sculpture of the Olympian Zeus) — they were constructed with strict steadi ness and harmony based on the golden mean. 1.1.3. A Way of Geometrical Construction of the Golden Section C The golden section arises often in geometry. From the Euclidean Ele D ments, we know the fol lowing form of geometric construction of the gold en section by using only a pair of compasses and a A B ruler (Fig. 1.3). E Construct a right tri angle ABC with the sides Figure 1.3. The geometric construction of the golden section AB=1 and AC=1/2. Then, 2 according to the Pythagorean Theorem, we have: CB = 1 + (1 / 2) = 5 / 2 . By drawing the arc AD with the center at the point C before its intersec tion with the segment CB at the point D, we obtain the segment ( ) BD = CB − CD = 5 − 1 / 2 = τ−1 . (1.7) By drawing the arc DE with the center at the point B before its intersection with the segment AB at the point E, we obtain a division of the segment AB at the point E by the golden section because AB EB (1.8) = = τ or AB = 1 = EB + AE = τ −1 + τ −2 . EB AE Thus, the simple right triangle with leg ratio of 1:2, well known to ancient science, could be the basis for the discoveries of the Pythagorean Theorem, the Golden Section and Incommensurable Line Segments — the three great mathe matical discoveries attributed to Pythagoras. 1.1.4. A “Double” Square Many mathematical regularities, we can say, did “lie on the surface,” and they needed only to be seen by a person with analytical, logical thinking that Chapter 1 The Golden Section 7 was inherent in antique philosophers and mathematicians. It is possible that ancient mathematicians could find the golden proportion by investigating the socalled elementary rectangle with the side ratio of 1:2. This is named a “double” or “twoadjacent” square because it consists of two squares (Fig. 1.4). Let AB = 1 and AD = 2 . Then, if we A D calculate the diagonal AC of the “dou ble” square, then according to the Pythagorean Theorem we get: 5 AC = 5. If we take the ratio of the sum AB + AC relative to the larger side АD B C of the “double” square, we come to the Figure 1.4. A rectangle with a side ratio golden mean, because: of 2:1 (a “double” square) ( ) (1.9) (AB+AC)/AD = 1 + 5 / 2 . It is paradoxical that the Pythagorean Theorem is very wellknown to each schoolboy while the golden mean is familiar to very few. The main pur pose of this book is to tell about this wonderful discovery of antique science to everyone who preserved the feeling to be surprised and admired, and we will show to modern scientists the far, not trivial, applications of the golden mean in many fields of modern science. We will tell our readers about this unique mathematical discovery, which during the past millenia has attracted the at tention of outstanding scientists, mathematicians and philosophers of the past, like: Pythagoras, Plato, Euclid, Leonardo da Vinci, Luca Pacioli, Johannes Kepler, Zeising, Florensky, Ghyka, Corbusier, Eisenstein, American mathe matician Verner Hogatt — founder of the Fibonacci Association, and the Great scientist Alan Turing — a founder of modern computer science. 1.2. Algebraic Properties of the Golden Mean 1.2.1. Remarkable Identities of the Golden Mean What a “miracle” of nature and mathematics is the golden mean? Why does our interest in it not wither, but, on the contrary, increases with each century? To ponder this issue, we suggest that our readers focus all of their mathematical knowledge and plunge into this fascinating aspect of the world of mathematics. Only then there is the possibility of enjoying and under THE MATHEMATICS OF HARMONY Alexey Stakhov 8 standing the wonderful mathematical properties, beauty and harmony of this unique phenomenon — the golden mean. Let us start from the algebraic properties of the golden mean. It follows from (1.4) — which is very simple and nevertheless a rather surprising property of the golden mean. If we substitute the root τ (the golden mean) for x in Eq. (1.4), then we will get the following remarkable identity for the golden mean: τ2 = τ + 1. (1.10) Let us be convinced that the identity (1.10) is valid. For this purpose, it is necessary to carry out elementary mathematical transformations over the left hand and righthand parts of the identity (1.10) and to prove that they coincide. In fact, we have for the righthand part: 1+ 5 3+ 5 τ +1 = +1 = . 2 2 On the other hand, 2 1+ 5 1+ 2 5 + 5 3 + 5 2 = = , τ = 2 4 2 from whence the validity of the identity (1.10) follows. If we divide all terms of the identity (1.10) by τ we come to the following expression for τ: 1 τ = 1+ . (1.11) τ This can be represented in the following form: 1 τ −1 = . (1.12) τ Now, analyze the identity (1.12). It is well known that any number а has its own inverse number 1/a . For example, the fraction 0.1 is an inverse num ber to 10. A traditional algorithm to get the inverse number 1/a from the initial number а consists of the division of the number 1 by the number а. In general case, this is a very complex procedure. Try, for example, to get the inverse number of a = 357821093572. This can be fulfilled only by the use of modern computer. Consider the golden mean τ = 1 + 5 / 2, which is an irrational number. How can we get the inverse number 1/τ? The formula (1.12) gives a very simple answer to this question. To solve this problem, it is enough to subtract 1 from the golden mean τ. In fact, on the one hand, 1 2 2(1 − 5 ) 2(1 − 5 ) 5 −1 = = = 2 = . 2 2 τ 1 + 5 (1 + 5 )(1 − 5 ) (1) − ( 5 ) ( ) Chapter 1 The Golden Section 9 While on the other hand, as follows from (1.12), the inverse number 1/τ can be found in the following manner: 1 1+ 5 5 −1 = τ −1 = −1 = . τ 2 2 However, we will get greater “aesthetic pleasure” if we carry out the following transformations of the identity (1.10). Multiply both parts of the identity (1.10) by τ, and then divide them by τ 2. The outcome we get is two new identities: (1.13) τ3 = τ2 + τ and τ = 1 + τ −1. (1.14) If we continue to multiply both parts of the identity (1.13) by τ, then divide both parts of the identity (1.14) by τ 2, and continue this procedure ad infinitum, we will come to the following graceful identity that connects the adjacent degrees of the golden mean: (1.15) τn = τn −1 + τn −2 , where n is an integer taking its values from the set {0,±1,±2,±3,…}. The identity (1.15) can be expressed verbally as follows: “Any member of the golden series (golden powers) is the sum of the previous two golden powers.” This property of the golden mean is truly unique! It is very difficult to believe that the following identity is “absolutely true”: 100 99 98 100 99 97 ⎛ 1+ 5 ⎞ ⎛ 1+ 5 ⎞ ⎛ 1+ 5 ⎞ τ =⎜ +⎜ =⎜ (1.16) ⎟ ⎟ ⎜ 2 ⎟⎟ , ⎜ 2 ⎟ ⎜ 2 ⎟ ⎠ ⎠ ⎝ ⎠ ⎝ ⎝ however, its validity follows from the validity of the general identity (1.15). Moreover, the following identity is also true: 100 96 ⎛ 1+ 5 ⎞ ⎛ 1+ 5 ⎞ ⎛ 1+ 5 ⎞ ⎛ 1+ 5 ⎞ +⎜ +⎜ =⎜ τ =⎜ (1.17) ⎜ 2 ⎟⎟ . ⎜ 2 ⎟⎟ ⎜ 2 ⎟⎟ ⎜ 2 ⎟⎟ ⎠ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎝ There are an infinite number of identities similar to (1.17), and all of them follow from the general identity (1.15). 100 1.2.2. The “Golden” Geometric Progression Consider the following sequence of golden mean degrees: − n −1 ..., τ− n , τ ( ) ,..., τ−2 , τ−1 , τ0 = 1, τ1 , τ2 ,..., τn −1 , τn ,... . { } (1.18) The sequence (1.18) has very interesting properties. On the one hand, the sequence (1.18) is a geometric progression because each of its element is Alexey Stakhov THE MATHEMATICS OF HARMONY 10 equal to the preceding one multiplied by the number τ, a Denominator or Constant Multiplier in the geometric progression (1.18), that is, τn = τ × τn −1 . (1.19) On the other hand, according to (1.15), the sequence (1.18) has the prop erty of “additivity” because each element is equal to the sum of the two pre ceding elements. Note, that the property (1.15) is characteristic only for the geometrical progression with the denominator τ, and such geometrical pro gression is named the Golden Progression, Golden Series or Golden Powers. In this connection, any Logarithmic Spiral corresponds to a certain geo metrical progression of the type (1.18), the opinion of many researchers is that the property (1.15) distinguishes the golden progression (1.18) among other geometrical progressions, and this fact is a reason for the widespread prevalence of the golden logarithmic spiral in forms and structures of nature. 1.2.3. A Representation of the Golden Mean in the Form of Continued Fraction Let us now prove another astonishing property of the golden mean based on the identity (1.11). If we substitute the expression 1+1/τ for t in the righthand part of (1.11), then we get the representation t in the form of the following continued fraction 1 τ = 1+ . 1 1+ τ If we continue such substitution ad infinitum, we get the continued frac tion of the following form: 1 (1.20) τ = 1+ . 1 1+ 1 1+ 1 1+ 1 + ... A representation of (1.20) in mathematics is called a Continued or Chain fraction. Note that the theory of continued fractions is one of the most impor tant topics of modern mathematics. 1.2.4. A Representation of the Golden Mean in “Radicals” Consider the identity (1.10). If we extract a square root from the right hand and lefthand parts of the identity (1.10), then we will get the following representation for τ: τ = 1 + τ. (1.21) Chapter 1 The Golden Section 11 If we substitute the expression 1+ τ for τ in the righthand part of (1.21), then we get the following representation for τ: (1.22) τ = 1+ 1+ τ . If we continue such substitution ad infinitum, we will get another re markable representation of the golden mean τ in “radicals”: τ = 1 + 1 + 1 + 1 + ... . (1.23) Every mathematician intuitively aspires to express mathematical results in the simplest and most compact form, and if he discovers such an expression, it gives him “aesthetic pleasure.” In this respect, mathematical creative work (aspiration for “aesthetic” expression of mathematical results), is similar to the creative activity of a composer or poet, because their main aim is to find perfect musical or poetic forms that give rise to aesthetic pleasure. Note that the formulas (1.20) and (1.23) give us an aesthetic pleasure, arousing a feeling of rhythm and harmony when we begin to think of the infinite repeatability of the same simple mathematical elements in formulas for τ. 1.3. The Algebraic Equation of the Golden Mean 1.3.1. Polynomials and Equations Since ancient times, mathematicians paid special attention to the study of polynomials and the solutions of algebraic equations, and this important math ematical problem promoted the development of algebra. As is well known, a Polynomial of nth degree is represented by the following expression: (1.24) an x n + ... + a2 x 2 + a1 x + a0 , where In other words, a polynomial is the sum of the integervalued degrees taken with certain coefficients. For example, a decimal notation of a number is, in essence, some representation of this number in the form of a polynomial of the number 10, for instance, 365 = 3 × 102 + 6 × (10 ) + 5 . If x is a variable, then a polynomial gives any Polynomial Function, the range of which coincides with the range of the variable x. The polynomials of the first, second, third, and fourth degrees are called Linear, Quadratic, Cubic, and Biquadrate polynomials, respectively. The Algebraic Equation (in the standard form) is a statement written with algebraic designations that some polynomial function is equal to zero for some values of a variable x. For example, ( ) Alexey Stakhov THE MATHEMATICS OF HARMONY 12 x 2 − 5 x + 6 = 0 is an algebraic equation. The values of the variable, for which the polynomial becomes equal to 0, are named roots of this polynomial. For example, the polynomial x 2 − 5 x + 6 has two roots, 2 and 3, because 22 − 5 × 2 + 6 = 0 and 32 − 5 × 3 + 6 = 0. Note that in the polynomial x 2 − 5 x + 6 the variable x represents an arbitrary number from the range of the given polynomial function, but the alge braic equation x 2 − 5 x + 6 = 0 , in contrast, means that the variable x can only be the number 2 or 3, the roots of this algebraic equation. Algebraic equations have always served as a powerful means to solutions of practical problems. The exact language of mathematics has allowed mathemati cians to simply express the facts and relations, which, when presented in ordi nary language, can seem confusing and complicated. The unknown values, de noted by some algebraic symbols, for example x, can be found if we formulate the problem in mathematical language through the form of equations. Methods of an equation’s solution are the basis for the Equation Theory in mathematics. Now, recall the basic data of linear and square algebraic equations that are wellknown to us from secondary school. The linear equation, in general form, can be written as ax + b = 0, where a and b are some numbers and a ≠ 0. This equation has one solution: x = −b / a ; that is, the linear equation has only one root. The quadratic equation has the following form: (1.25) ax 2 + bx + c = 0, where a ≠ 0. The rules for the solution of algebraic equations of the first and second degree were wellknown in antiquity. For example, as we recall from second ary school mathematics, we know the following formula for the roots of the quadratic equation (1.25): −b ± b2 − 4ac (1.26) . 2a Remember that the values of the roots depend on the Determinant D of the quadratic equation (1.25) — which is defined by the following formula: x1,2 = D = b2 − 4ac. (1.27) If the determinant D is positive, the formula (1.26) gives exactly two real roots. If D = 0, then x = −b / 2a , and we say that Eq. (1.25) has two equal roots. If the determinant D is negative, we have to introduce an imaginary unit i, which is defined by i = −1, and for this case, both of these roots are complex. 1.3.2. Quadratic Irrationals The discovery of irrational numbers is the greatest mathematical discovery of Greek mathematics. This discovery was made by Pythagoreans in the 5th century Chapter 1 The Golden Section 13 B.C. during the investigation of the ratio of two geometric segments: a diagonal and a side of a square. By using Reductio ad Absurdum, that is, the Indirect Method of Con tradiction, the Pythagoreans proved that the ratio of the diagonal to the side of the square, which is equal to 2 , cannot be represented in the form of the ratio of two natural numbers. Such geometric segments were named Incommensurable, and the numbers, which express similar ratios, were named Irrational Numbers. The discovery of Incommensurable Segments resulted in the first crisis of the foundations of mathematics, and, ultimately, became a turning point in the development of mathematics. One legend says that in honor of this dis covery, Pythagoras had carried out a “hecatomb,” namely, he had sacrificed 100 bulls to the gods. In the beginning, the Pythagoreans attempted to keep the new discovery a secret. According to legend, Hippias, one of the Pythagore ans, broke the oath and divulged the secret of this discovery. Later, he per ished during a shipwreck that the Pythagoreans considered to be a punish ment from the gods for the disclosure the discovery’s secret. The influence of this discovery on the development of science can be compared to the discov ery of Lobachevsky’s geometry in the first half of the 19th century, and with the discovery of Einstein’s Relativity Theory at the beginning of the 20th century. Thanks to this discovery, mathematics received an entirely new math ematical concept — the concept of Irrational Numbers. The title “rational” is from the Latin word “ratio,” which is a translation of the Greek word “logos.” The numbers, which can be represented as the ratios of two integers, were called “rational.” In contrast to rational numbers, those numbers, which express the ratios of the incommensurable line segments, had been named irrational (from the Greek word “alogos”). Unlike a linear equation, a quadratic equation of the kind (1.25) with ra tional coefficients can have irrational roots called quadratic irrationals. Book 10 of The Elements is devoted to the study of the quadratic irrationals. Scientists of India, the Middle East, and many European mathematicians of the Middle Ages studied quadratic irrationals. It was proved that any number of the kind N for any integer N, which is not a full square, is an irrational number as well as a number 3 N , where N is not a cube, etc. As they say, similar irrationals can be represented in Radicals. In 16th century, Italian mathemati cians Tartalja, Cardano, and Ferrari found a general formula for the roots of the cubic and biquadrate algebraic equations that represented their irrational roots in radicals. Despite the fact that attempts to find general formulas for the roots of the algebraic equations of the fifth and more degrees appeared unsuccessful, this research direction resulted in new mathematical discoveries, which were connected with the mathematical works of Niels Abel and Evariste Galois. Alexey Stakhov THE MATHEMATICS OF HARMONY 14 1.3.3. Poor “Studious” Niels Abel The history of science and mathematics contains many tragic pages. One of them is the life and scientific work of the Norwe gian mathematician Niels Abel, whose mathematical research in volves algebra and solutions to algebraic equations. For a long time, mathematicians assumed that an irratio nal root of any algebraic equation with rational factors can be expressed in radicals. However, in the 19th century, Abel proved that the irrational roots of the general equations of Niels Henrik Abel (18021829) the 5th and more degrees cannot be expressed in radicals. On the big square in the city of Oslo, the capital of Norway, a majestic monument rises up above the cityscape. On the granite stone of the monu ment, a young man is carved with a spiritualized face struggling against two disgusting monsters. The monument is of the wellknown Norwegian mathe matician Niels Henrik Abel. What do these monsters symbolize? Some math ematicians joke that these monsters represent the equations of the fifth de gree and the elliptic functions conquered by Abel. Others consider that it is allegory: the sculptor wished to embody the social injustice with which Abel struggled all his life. The sad life of this Norwegian mathematical genius is typical for scientific geniusesnot just from his country or of his time. Unfortunately, the destiny of many mathematical geniuses unfolds tragically. It is connected with the fact that many of great mathematicians are often not understood, and their stunning mathematical discoveries are not recognized by their contemporar ies, unfortunately, they are only recognized some time after their death. Abel was born in 1802 in Northwest Norway in the small fishing town of Finnej, where there were neither mathematicians, nor mathematical textbooks available to him. Very little is known about the first years of his childhood. He was enrolled in school in Oslo at the age of thirteen. Abel studied lightly and obtained good marks excelling in mathematics. He also liked to play chess and visit the theatre. However, within three years of school, a sudden change opened up his world. Instead of a cruel mathematics teacher who beat pupils, the new mathe matics teacher, Holmboe, arrived at the school. He knew his subject very well, and he spiked the interest of his pupils. Holmboe gave each pupil the opportunity to act independently, and encouraged them to take the first steps paid special attention towards a more fulfilling mathematics education. Soon, Abel not only took a great interest in mathematics, but also found that he was capable of solv ing mathematical problems of which other people were incapable. Chapter 1 The Golden Section 15 Abel’s family lived in humiliating pov erty so he was attending a school free of charge. In 1820, Abel’s father died, and the family remained without any financial means. Their situation was desperate. Niels thought about returning to his native city and searching for a job, but some of his pro fessors paid attention to the young man’s talent and helped him to enroll in the local university. They managed to obtain grants for Niels Abel to preserve his unique math ematical talent in science, and for travel ing abroad. Abel’s visits to Berlin, Paris, and to other large mathematical centers of that time contributed greatly to his future mathematical works. Unfortunately, the young mathematician’s discoveries which A monument of Niels Abel in Oslo surpassed much the science of the time resulted in a lack of understanding and un derestimation by his contemporaries. Abroad, as well as in his native land, Abel ex perienced great financial difficulties and a constant feeling of intolerable loneliness. His attempts to get scientific recognition were unsuccessful. He sent his works to the Parisian Academy where they were forwarded to the French mathematician, Cauchy. Unfortunately, these works were lost. Also his letter to the German math ematician Gauss was left unanswered. The young mathematician, who made a revolution in mathematics, re turned home to remain the same poor and unknown “studious” Abel. He could not find a job. Being ill by tuberculosis, and “poor as a church mouse,” by his own words, Abel, in a condition of the cheerless melancholy, died on April 6, 1829, at 26. The proof of insolvability in radicals of the algebraic equations of the fifth and greater degrees was his most important mathematical discov ery, made at just 22. In the opinion of the mathematician Ermit, “Abel left such a rich heritage for mathematicians that they will develop it dur ing the next 500 years.” There is a custom, to which new results and discoveries are named after the scientists who made them. Now, anyone holding a book on higher mathe matics can see that Abel’s name is immortalized in various areas of mathemat ics. There are several theorems and mathematical results bearing Abel’s name: Alexey Stakhov THE MATHEMATICS OF HARMONY 16 Abelian integrals, Abelian equations, Abelian groups, Abelian formulas, and Abelian transformations. Abel would be surprised to learn that his work has had a profound influence upon the development of mathematics. 1.3.4. Mathematician and Revolutionist Evariste Galois The destiny of one more genius, the French mathematician Evariste Galois, is not less tragic. He was born on October 26, 1811, in the small town of Bourg laReine near Paris, France. For the first twelve years of his life his mother, who was a fluent reader of Latin and classical literature, was responsible for his education. In October 1823, he entered the Lycee Lou isleGrand. At LouisleGrand, Galois enrolled in the mathe matics class of Louis Richard. He worked more and more on his own researches and less and less on his schoolwork. He stud ied Legendre’s Geometrie and the treatises of Lagrange. His teacher Richard reported, “This student works only in the highest realms of mathematics.” In April 1829 Galois published his first mathematics paper on continued fractions in the An nales de Mathematiques. Later he submitted articles on the Evariste Galois (18111832) algebraic solution of equations to the Academie des Sciences. Cauchy was appointed as referee of Galois’ paper. On Decem ber 29, 1829 Galois has received his Baccalaureate degree. In the development of Abel’s work that he learned from Bulletin de Ferussac, and by the Cauchy’s advice, he submitted a new article On the Condition that an Equation be Soluble by Radi cals in February 1830. The paper was sent to Fourier, the secretary of the Paris Academy, to be considered for the Grand Prize in mathematics. Since Fourier died in April 1830, Galois’ paper was never found and considered for the prize. Further Galois’ works on the theory of elliptic functions and Abelian integrals were initiated by Abel and Jacobi’s works. With support from Jacques Sturm, he published three papers in Bulletin de Ferussac in April 1830. Galois enthusiastically participated in revolutionary activity, and even tually found himself in a prison for several times. In March 1832, being in a prison, he felt ill with cholera and was sent to the pension Sieur Faultrier where he fell in love with StephanieFelice du Motel, the daughter of the resident physician. In May 1832 his stormy life finished: he was killed in a duel. The reason for the duel was not clear but certainly linked with Stephanie. Before the duel, he wrote a resume of his mathematical discover ies and transferred his notes to one of his friends, requesting him to pass it on Chapter 1 The Golden Section 17 to one of the leading mathematicians. The note ended with the following words: “You will publicly ask Jacobi or Gauss to give a conclusion not about validity, but about the value of these theorems. Then, I hope, people, who can decipher all this mess, will be found.” From what is known, Galois’ letter reached neither Jacobi nor Gauss. The mathematical community only learned of Galois’ works in 1846, when the French mathematician Liouville published most of Galois’ works in his mathematical journal. All of Galois’ works were presented in a small format of 60 pages. These works contained a presenta tion of group theory — now considered to be the “key” theory of modern algebra and geometry, the first classification of irrationals defined by alge braic equations. This doctrine is now referred to as Galois Theory. 1.3.5. The “Golden”Algebraic Equations of nth Degree And now, after such fundamental mathematical training, we again turn to the equation of the golden proportion (1.4). Clearly, it is the quadratic algebraic equation of the type (1.25) with the factors: (1.28) a = 1; b = −1; c = −1. By using (1.27) and (1.28), we can calculate the determinant of Eq. (1.4): D = 5. It follows from here that Eq. (1.4) has two real roots: 1+ 5 1 1− 5 . and x2 = − = (1.29) 2 τ 2 The root x1 coincides with the golden mean τ. Typically, the main problem of the theory of algebraic equations is to find their roots. We know the elementary algebraic equation (1.4) with the root equal to the golden mean. With this equation, the following question arises: are there algebraic equations of the highest degrees with the root equal to the golden mean? And if so, what do they look like? To answer this ques tion we will use the following reasoning concerning the initial equation of the golden mean (1.4). Multiply both parts of Eq. (1.4) by x; as a result we get the following equality: x1 = τ = x 3 = x 2 + x. (1.30) Equation (1.4) can be written in the following form: x = x 2 − 1. If we substitute x2 1 for x in the expression (1.30), we get the following THE MATHEMATICS OF HARMONY Alexey Stakhov 18 algebraic equation of the third degree: (1.31) x 3 = 2 x 2 − 1. 2 On the other hand, if we substitute x +1 for x in (1.30), then we get one more equation of the third degree: (1.32) x 3 = 2 x + 1. Thus, we have obtained two new algebraic equations of the third de gree. Now, prove that the golden mean is a root of Eq. (1.31). With this end, we substitute the golden mean τ = 1 + 5 / 2 for x to the lefthand and righthand parts of Eq. (1.31) and show that both parts of this equa tion coincide. In fact, by using the identity (1.15), we can obtain for the lefthand part: ( ) 3 + 5 1+ 5 4 + 2 5 + = = 2 + 5. 2 2 2 We can also obtain for the righthand part of this equation: τ3 = τ2 + τ = 3+ 5 − 1 = 2 + 5. 2 Hence, the golden mean τ is, in fact, a root of Eq. (1.31), that is, Eq. (1.31) is a golden one. By analogy, we can prove that Eq. (1.32) is a golden one as well. Let us derive the golden algebraic equation of the fourth degree, for this purpose we multiply both parts of the equality (1.30) by x; as a result we will get the following equality: 2τ 2 − 1 = 2 × (1.33) x4 = x3 + x2. 2 We then use the expression (1.4) for x and the expressions (1.31) or (1.32) 3 for x . If we substitute them into the expression (1.33), we obtain two new golden algebraic equations of the fourth degree: x 4 = 3x 2 −1 (1.34) (1.35) x 4 = 3 x + 2. The analysis of Eq. (1.35) resulted in unexpected result. It is found, that this equation describes the energetic state of the butadiene molecule, a valu able chemical substance used in the production of rubber. American scientist Richard Feynman, Nobel Prize Laureate, expressed his enthusiasm concern ing Eq. (1.35) in the following words: “What miracles exist in mathematics! According to my theory, the golden proportion of the ancient Greeks gives the minimal energy condition of the butadiene molecule.” This fact, at once, raises our interest in the golden equations of the higher Chapter 1 The Golden Section 19 degrees, which, probably, describe energy conditions of the molecules of other chemical substances. These equations can be obtained, if we will consistently consider the equalities of the kind x n = x n −1 + x n −2 . As an example, we can derive the following golden equations of the higher degrees: x 5 = 5 x 2 − 2 = 5 x + 3; x 6 = 8 x 2 − 3 = 8 x + 5; x 6 = 13 x 2 − 5 = 13 x + 8. The analysis of these equations demonstrates that Fibonacci numbers (1, 1, 3, 5, 8, 13, 21, 34, 55) — which are addressed in Chapter 2 — are the numerical factors in the righthand part of these equations. It is easy to show, that in the general case, the golden algebraic equation of the nth degree can be expressed in the following form: x n = Fn x 2 − Fn −2 = Fn x + Fn −1 , (1.36) where Fn , Fn −1 , Fn −2 are Fibonacci numbers. Note, once again, that the main mathematical property of all equations of the kind (1.36) is the fact that all of them have a common root — the golden mean. Thus, our simple reasoning resulted in a small mathematical discovery: we have found an infinite number of new golden algebraic equations given by (1.36) — which have the golden mean as a common root. If we substitute in Eq. (1.36) its root τ = 1 + 5 / 2 for x, then we obtain the following remarkable identities that connect the golden mean with Fi bonacci numbers: (1.37) τn = Fn τ2 − Fn −2 = Fn τ + Fn −1 . Below we show that, for instance, the18th, 19th, and 20th Fibonacci num bers are equal to the following: (1.38) F18 = 2584, F19 = 4181, F20 = 6765. Then, taking into consideration (1.38) and using (1.36), we can write the following algebraic equations of the 20th degree: (1.39) x 20 = 6765 x 2 − 2584 ( ) (1.40) x 20 = 6765 x + 4181. τ = 1 + 5 / 2 is a root of It is difficult to imagine that the golden mean the equations (1.39) and (1.40). However, this fact follows from the theory of golden algebraic equations given by (1.36), and by looking at the equations (1.36), (1.39) and (1.40). Once again, we are convinced of the greatness of mathematics — which allows us to express the complex scientific information in such compact form. ( ) Alexey Stakhov THE MATHEMATICS OF HARMONY 20 1.4. The Golden Rectangles and the Golden Brick 1.4.1. A Golden Rectangle with a Side Ratio of τ Many different definitions A of the golden rectangles exist. First, we will study the golden rectangle with a side ratio τ (Fig. 1.5). A rectangle in Fig. 1.5 is a golden one because the ratio of its larger side to the smaller side is equal to the gold en mean, that is: E B O G I H 1+ 5 F D C . Figure 1.5. Golden rectangle with a side ratio of τ 2 Consider the golden rectangle with the sides AB = τ and BC = 1 . We can find points E and F on the line segments АВ and DC, which divide the seg ments АВ and DC in the golden section. It is clear that if AE = DF = 1 , then EB = AB − AE = τ − 1 = 1 / τ . Now, we draw the line segment EF, which is called a Golden Line. It is clear that the golden line EF divides the golden rectangle ABCD into two new rectangles AEFD and EBCF. It follows from the above geometric con siderations that the rectangle AEFD is a square. Now, let us consider the rectangle EBCF. As its big side BC = 1 and its small side EB = 1 / τ , hence their ratio BC : EB = τ . It means that the rectan gle EBCF is a golden one also. Thus, the golden line EF divides the initial golden rectangle ABCD into a square AEFD and a new golden rectangle EBCF. Now, draw the diagonals DB and EC of the golden rectangles ABCD and EBCF. It follows from the similarity of the triangles ABD, FEC, BCE that the point G divides the diagonal DB in the golden section. Then, we draw a new golden line GH in the golden rectangle EBCF. It is clear that the golden line GH divides the golden rectangle EBCF into a square GHCF and a new golden rect angle EBHG. By repeating this procedure infinitely, we will get an infinite se quence of squares and golden rectangles converging in the limit to the point O. Note that such endless repetition of the same geometric figures, that is, squares and golden rectangles, gives rise to aesthetic feelings of harmo ny and beauty. It is considered, that this fact is a reason why many rectan AB : BC = τ = Chapter 1 gular forms (match boxes, lighters, books, suitcases) frequently are golden rectan gles. Below we will discuss the applica tions of the golden rectangle to architec ture and painting. For example, we widely use credit cards in our daily life, however we do not pay attention that in many cases our cred it cards have a form that is, or approxi mates to the golden rectangle (Fig. 1.6). The Golden Section 21 Figure 1.6. A credit card has a form of the golden rectangle 1.4.2. A Golden Rectangle with a Side Ratio of τ 2 We look back to Theorem II.11 of The Elements of Euclid. Consider, once again, a division in the extreme and mean ratio represented in Fig. 1.3, and according to Theorem II.11, we should construct the line segments АВ, АС, and СВ, which form the golden mean (Fig. 1.2), a rectangle for the condition (1.2). We name the rectangle a Euclidean rectangle. If we designate the lengths of the line segments АВ, АС and СВ as AB = a, AC = b and CB = c, then, we can rewrite the expression (1.2) as follows: a × c = b2 . (1.41) Taking into account (1.41), we can represent the Euclidean rectangle as is shown in Fig. 1.7. c = τ2 = 0.382 It follows from Fig. 1.7 that a ratio of the larger side to the smaller one in the Euclidean rectangle is equal to the ratio of the initial seg a =1 Figure 1.7. The Euclidean rectangle ment to the smaller segment in Theorem II.11 of The Elements here, according to (1.41), its area is equal to a square of the larger segment. If we take the segment of the length 1 as the initial segment ( a = 1) , then the lengths of the larger segment (b) and smaller one (c) in the DEMR will always be a proper fraction, and the expression (1.41) can be written in the following form: c = b2 . (1.42) The following formulation of Theorem II.11 for the case of the line seg ment of the length 1 comes directly from (1.42). Theorem II.11 for the segment of the length 1. We divide the line seg ment of the length 1 into two unequal parts in golden ratio; thus, the length of the smaller line segment will be equal to the square of the larger line segment. Alexey Stakhov THE MATHEMATICS OF HARMONY 22 If the initial segment АВ in Fig. 1.1 will be the line segment of length 1, −1 that is, AB = 1 , then the larger segment AC = τ , and the smaller segment CB = τ−2 . Taking into account this fact, we can say that the Euclidean rectan gle in Fig. 1.7 is a new golden rectangle with a side ratio AB : CB = τ2 . It fol lows from this consideration that Euclid, in his Theorem II.11, not only for mulated the DEMR (the golden section), but he also discovered a new golden rectangle with a side ratio τ2 . Note that Theorem II.11, for the segment with length 1, presents the following wellknown property of the golden mean: 1 = τ–1 + τ–2 = 0.618 + 0.382. (1.43) The identity (1.43) is a partial case of the identity (1.6), and it expresses the famous “Principle of the Golden Proportion” that has been widely used in human culture since antiquity. 1.4.3. The “Golden” Brick of Gothic Architecture Again, we will use the Euclidean rectangle in Fig. 1.8, where the larger side is AB = τ and the smaller side is AD = τ−1 . Draw the Euclidean rectangle in Fig. 1.7 with a diagonal DB. A B τ 1 3 τ D C Figure 1.8. A calculation of the diagonal in the Euclidean rectangle By using the Pythagorean Theorem, we can write: DB = BC 2 + DC 2 = τ2 + τ–2. Calculate numerical values for τ2 and τ−2 . In fact, we have: 2 (1.44) 2 1+ 5 1+ 2 5 + 5 3 + 5 = = τ = . 2 4 2 On the other hand, we have: 2 (1.45) 2 τ −2 5 −1 5 − 2 5 +1 3 − 5 = = = . 2 4 2 (1.46) Chapter 1 The Golden Section 23 Taking into account (1.45) and (1.46), we can rewrite the expression (1.44) as follows: DB2 = 3. (1.47) It comes from (1.47) that DB = 3. (1.48) As is shown in [26], the Euclidean rectangle in Fig. 1.8, together with the classical golden rectangle in Fig. 1.5, can be used for the construction of a special kind of rectangular parallelepiped — also called a Golden Brick (Fig. 1.9). Consider this parallelepiped for the case AB = DC = 1 . The faces of the golden brick in F Fig. 1.9 are the golden rectan gles with geometric ratios G based on the golden 2 mean. The face 1 ABCD is a clas E τ sical golden H rectangle with A 3 the side ratio of B 1 τ. This means τ that the edge is AD = τ−1 . The τ face ABGF is D also a classical 1 C golden rectan gle with the Figure 1.9. The golden brick side ratio of τ. This means that the edge AF = τ . Finally, the rectangle BCHG is the Euclid ean rectangle with the side ratio of τ2 (Fig. 1.7). This means that the edges are BG = τ and BC = τ−1 . Note that the diagonal CG = 3. By using the Pythagorean Theorem, we can calculate the diagonal CF of the golden brick: CF = FG 2 + CG 2 = 1 + ( 3 )2 = 2. In the book [26], evidence is suggested that the golden bricks were wide ly used in the Gothic castles as a form of basic building blocks. In [26], the hypothesis is drawn that the surprising strength of the Gothic castles is bound up with the use of the golden bricks during the construction of architectural monuments of gothic style. Alexey Stakhov THE MATHEMATICS OF HARMONY 24 1.5. Decagon: Connection of the Golden Mean to the Number π The quantity of irrational (incommensurable) numbers is infinite. Howev er, some of these numbers take a special place in the history of mathematics — specifically, in the history of material and spiritual culture. Their significance involves the fact that they express proportions and relations — which have uni versal character — and often appear in the most unexpected places. The irratio nal number 2 — which is equal to the ratio of the diagonal to the side of a square — is the first of them. Recall that the discovery of the socalled “incom mensurable line segments” is connected directly with this number. This discov ery caused the most dramatic period in ancient mathematics. It resulted in the development of the theory of irrational numbers and ultimately to the creation of modern “continuous” mathematics. Two mathematical constants — the number π, which expresses the ratio of a circle length to its diameter, and “Euler’s Number” е, the base of natural logarithms are the next important irrational numbers. Their significance consists in the fact that they originate the main classes of the “elementary functions”: trigonometric functions (the number π) and also exponential func tion ex, logarithmic function logex, and lastly hyperbolic functions (the num ber е). The numbers π and е, that is, two of the most important constants of mathematical analysis, are connected by the following astonishing formula: 1 + e ix = 1, (1.49) where i = −1 is an imaginary unit. This identity is not recognized by everyone at first sight. However, in con a10 sidering this identity, which connects among themselves the fundamental R mathematical constants, the numbers 36° π and e, it is difficult to refrain from pondering the mystical character of this mathematical formula (1.49). The golden mean τ refers also to a category of the fundamental mathe Figure 1.10. Decagon matical constants, but this poses a question. Is there some connection be tween these mathematical constants, for example, between the numbers π and τ? The analysis of a regular Decagon (Fig. 1.10) gives the intriguing an swer to this question. Chapter 1 The Golden Section 25 Consider a circle with a radius R and a decagon inscribed inside the circle (Fig. 1.9). It is known from geometry that the side a10 of the decagon is con nected to the radius R by the following formula: a10 = 2Rsin18°. (1.50) If we carry out some trigonometric transformations by using the for mulas, wellknown from secondary school trigonometry, we obtain the fol lowing outcomes: (1) The side of the decagon inscribed inside the circle with radius R is equal to a large part of the radius R divided in the golden section, that is: a10 = R / τ. (2) The golden mean τ is connected with the number π by the following correlation: τ = 2cos36° = 2cos(π/5). (1.51) This formula, which is obtained as an outcome of mathematical analysis of the decagon, is one more confirmation of the fact that the golden mean to gether with the number π, shares the status of being amongst main mathe matical constants of nature. Consider the following from the area of nuclear physics. By studying the laws of the nuclear kernel, Byelorussian physicist Vasily Petrunenko, in the book The Golden Section of Quantum States and its Astronomical and Physical Manifestations (2005) [53], came to the following conclusion: a huge stabil ity of the nuclear kernel is due to the wave’s multiplicities of the golden mean underlying their organization. Thus, he proved that the regular decagon un derlies the structure of the nuclear kernel! 1.6. The Golden Right Triangle and the Golden Ellipse A 1.6.1. The Golden Right Triangle The right triangle based upon the golden sec tion is used widely in architecture. Consider the right triangle АВC with the side ratio AC : CB = τ (Fig. 1.11). If we designate x, y, and z as the side lengths of the right triangle АВС and also take into account that the z y C 51°50' x B Figure 1.11. The golden right triangle Alexey Stakhov THE MATHEMATICS OF HARMONY 26 ratio y : x = τ , then according to the Pythagorean Theorem, the length z of the side AB can be calculated by the formula: z = x 2 + y2 . (1.52) If we take x = 1, y = τ , then we have: z = 1 + τ = τ2 = τ. The right triangle with the side ratios τ : τ :1 is called a Golden Right Tri angle. Below, we will show that the golden right triangle is the main geometric idea of the Great Pyramid in Giza. 1.6.2. The “Golden” Ellipse C The golden ellipse [26] is H G formed with the help of the two golden rhombi (ACBD and ICJD) inscribed into the ellipse (Fig. J 1.12). The golden rhombi ACBD I O A B and ICJD consist of four right tri angles of the kind OCB or OCJ, which are the golden right trian gles (Fig. 1.11). F E D Now, consider the basic geo metric correlations of the golden Figure 1.12. The golden ellipse ellipse in Fig. 1.12. Let a focal dis tance AB of the ellipse in Fig. 1.12 be equal to AB = 2 . The following correla tion follows from the ellipse definition: AC + CB = AG + GB. (1.53) The following correlations, which connect the sides of the golden right triangles OCB and OCJ, follow from the definition of the golden right triangle. As the smaller leg OB of the golden right triangle is equal to 1, that is, OB = 1 , it follows from the definition of the golden right triangle that the values for the hypotenuse and the larger leg of the triangle OCB are: CB = τ and OC = τ . (1.54) Let us consider the golden right triangle OCJ. Because its smaller leg OC = τ, then, according to the definition of the golden right triangle, we have: OJ = OC × τ = τ × τ = τ and CJ = τ × OC = τ × τ . (1.55) Chapter 1 The Golden Section 27 There is also the wellknown correlation that connects the halfaxes OI and ОС with the line segment OB of the ellipse in Fig. 1.12: (1.56) OC 2 = OI 2 − OB2 . We know that the correlation (1.56) is true for the figure in Fig. 1.11. In fact, if we substitute in (1.56) the numerical values of the line segments OI = τ, OC = τ and OB = 1, we can obtain: 2 τ = τ2 − 1 or τ = τ2 − 1 that is equivalent to the identity (1.6). This means that the correlation (1.56) is valid and the figure in Fig. 1.11 is an ellipse. In the opinion of the Polish scientist Jan Grzedzielski, author of the book EnergeticznoGeometryczny Kod Phzyrody [26], the golden ellipse can be used as a geometric model for the spreading of the light in optic crystals, that is, the geometrical correlations of the golden ellipse give optimal conditions for the attainability by photons of the focuses with minimal energetic losses. ( ) 1.7. The “Golden” Isosceles Triangles and Pentagon 1.7.1. Construction of the “Golden” Isosceles Triangle and Regular Pentagon in The Elements Theorem II.11 played an important role in A C B The Elements by Euclid. Using this theorem, Eu clid constructs the isosceles triangle and regu lar pentagon, which are then used for the con struction of the dodecahedron (see Chapter 3). First, we construct an isosceles triangle whose D angles at its base are double relative to the ver tex angle (Fig. 1.13). We can do this by drawing the line AB and then divide it in the golden sec Figure. 1.13. A geometric construc tion at the point C. Then, we draw a circle with tion of the golden isosceles triangle the cenetr A and the radius AB. We mark on the circle the point D so that AC=CD=BD. The triangle ABD has the property that its angles B and D, at the base BD, are double relative to its vertex angle A. Note that the vertex angle A=36° and the angles B=D=72°. As C is the golden section point, AC=CD=BD, and AD=AD, this means that the ratio of the hips AB and AD to the base BD of the isosceles triangle ABD is equal to the golden mean, that is: Alexey Stakhov THE MATHEMATICS OF HARMONY 28 AB AD = =τ BD BD The isosceles triangle with a golden ratio (1.57) is aptly called the Golden Isosceles Tri angle (Fig. 1.13). By using the triangle ABD, we draw a cir cle through A, B, and D (Fig. 1.14). Then we bisect the angle ADB by the line DE, which meets the circle at the point E. Note that the line DE passes through the point C, which di vides a line AB in the golden section. Similarly we can find a point F and then we can draw the pentagon AEBDF. (1.57) A E F C D B Figure 1.14. A geometric con struction of a regular pentagon 1.7.2. A Regular Pentagon Let us consider a regular pentagon presented in Fig. 1.15. If we draw in the regular pentagon all di A agonals, we obtain a pentagonal star called a Pentagram or Pentacle. Note that the word “pentagon” originates from the Greek word F G B pentagonon and the word “pentagram” origi E nates from the Greek word “pentagrammon” H (“pente” is five and “grammon” is a line). A L “pentagram” is a regular pentagon with gold K en isosceles triangles constructed on each side. C D It is proved that the crossing points of diago nals in the pentagon are always the points of Figure 1.15. Regular pentagon and pentagram the golden section. Thus, they form a new reg ular pentagon: FGHKL. In this new pentagon, we can draw diagonals at their crossing points to form another pentagon. This process can be continued infinitely. Thus, the pentagon ABCDE consists of an infinite number of the pentagons formed by the crossing points of diag onals. This infinite repetition of the same geometrical figures creates a feel ing of rhythm and harmony, which is perceived by our intellect. The regular pentagon in Fig. 1.15 is a rich source of golden sections. If we assume that a side of the inner pentagon FGHKL is equal to 1, that is, FG = GH = HK = KL = 1, then we get the following properties of the regular pentagon and pentagram ABCDE based on the golden section: Chapter 1 The Golden Section 29 1. All five triangles, AGF, BHG, CKH, DLK, and EFL, are golden isosceles triangles; here the lengths of the ten segments AG = AF = BH = BG = CK = CH = DL = DK = EF = EL = τ. 2. All five sides of the regular pentagon ABCDE are equal to τ2 , that is, AB = BC = CD = DE = EA = τ2 . 3. All line segments AH = GC = AL = DF = EK = LC = τ2 . 4. All five diagonals of the regular pentagon ABCDE are equal to τ3 , that is, AC = AD = BE = BD = EC = τ3 . 5. The lengths of the segments AC, AH, HC, and GH are in geometric progression because AC = τ3 , AH = τ2 , HC = τ, GH = τ0 = 1. Some interesting information about the pentagram (pentacle) is presented in Dan Brown’s book The Da Vinci Code. The fivepointed star is clearly a pre Christian symbol connected with the worship and dedication of nature. The ancient people divided the world into two halves, male and female. They had gods and goddesses maintaining the balance of forces. When a male’s and fe male’s beginnings are balanced, then harmony is reigning in the world. When the balance is broken, chaos appears. The pentacle symbolizes the female half of the Universe. Historians, who study religions, name this symbol “the sacred female beginning,” or “the Sacred Goddess.” The fivepointed star symbolizes Venus, the Goddess of love and beauty. The Goddess Venus and the planet Ve nus are the same. The Goddess takes her place in the night sky and is known under many names: Venus, East Star, Ishtar, and Astarte, all of them symboliz ing the powerful female beginnings connected with nature and mother earth. Over an eight year period, the planet Venus describes a pentacle on the big circle of the heavenly sphere. Ancient people noticed this unique phenom enon and they were so impressed, that Venus and its pentacle became symbols for perfection and beauty. Today, only a few people know that the Olympic Games follow the half cycle of Venus; fewer people know that the fivepointed star missed out on becoming a symbol for the Olympic Games, because of a last moment modification: the five acute ends of the star were replaced with the five rings. Ironically, the organizers apparently believed that their chosen symbol better reflected the spirit and harmony of the Olympic Games. 1.7.3. The Pentagon The Pentagon (Fig. 1.16) is the headquarters of the United States De partment of Defense and is located in Arlington, Virginia. It is now consid ered to be the symbol of the U.S. military. The Pentagon is often used met Alexey Stakhov THE MATHEMATICS OF HARMONY 30 Figure 1.16. The Pentagon onymically to refer to the Department of De fense — rather than the building itself. The Pentagon is the largestcapacity office build ing in the world, and it is one of the world’s largest buildings in terms of floor area. It houses approximately 23,000 military and ci vilian employees plus approximately an addi tional 3,000 nondefense support personnel. It has five sides, five floors above ground (plus two basement levels), and five ring corridors per floor with a total of 17.5 miles (28 km) of corridors. 1.7.4. The “Golden Cup” and the Golden Isosceles Triangle A The regular pentagon in Fig. 1.15 includes a num ber of remarkable figures used widely in works of art. In ancient art, the law of the golden cup (Fig. 1.17) B was widely known. It was used by the antiquity’s E sculptors and masters of golden things. The shaded part of the pentagram in Fig. 1.17 gives a schematic F representation of the golden cup. The regular pentagon in Fig. 1.15 consists of C D five golden isosceles triangles. Each golden tri Figure 1.17. The golden cup A angle has an acute angle A = 36° at the top and two acute angles B = D = 72° at the base of the triangle (Fig. 17). The main fea 36° ture of the golden triangle in Fig. 1.18 consists of the fact that the ratio of the hip АС=AD to the base DC is equal to the golden mean. In studying the pentagram in Fig. 1.15 and H the golden triangle in Fig. 1.18, the Pythagoreans admired the fact that the bisector DH coincides with the diagonal DB of the pentagon (Fig. H'' 72° 72° 1.15), and it divides the side АС in the golden section at the point H. Moreover, the new H' H''' C D golden triangle DHC arises. If we draw the bi Figure 1.18. The golden sector of the angle H to the point H', and con isosceles triangle tinue this process ad infinitum, we will get an Chapter 1 The Golden Section 31 infinite sequence of the golden triangles. Furthermore, the case of the golden rectan gle in Fig. 1.5 and the pentagon in Fig. 1.15, the infinite repetition of the same geo metric figure (the golden triangle) after the drawing of the next bisector causes an aesthetic feeling of rhythm and harmony. 1.7.5. Pentagonal Symmetry in Nature In nature, the forms based on pentagonal symmetry (starfishes, sea hedge hogs, flowers) are widespread. The flowers of a water lily, a dog rose, a haw thorn, a carnation, a pear, a cherry, an appletree, a wild strawberry, and many other plants, consist of fivepetals. Below, in Fig. 1.19, we can see examples of nature’s structures based on the pentagonal symmetry. The presence of five fingers on a human’s hand and, the five bone rudi ments on the paws of many animals are an additional confirmation of the wide spread occurrence of pentagonal forms in the morphology of flora and the bi ological world. Figure 1.19. Pentagonal symmetry in nature Alexey Stakhov THE MATHEMATICS OF HARMONY 32 1.8. The Golden Section and the Mysteries of Egyptian Culture 1.8.1. Phenomenon of Ancient Egypt In the early 20th century, in Saqqara (Egypt), archaeologists opened a crypt in which the Egyptian architect Hesire was buried. The wood panels covered by a magnificent thread were extracted from the crypt together with different material objects. In total, there were the 11 panels in the crypt; among them, only 5 of the panels were preserved; the remaining panels were completely de stroyed by moisture inside the crypt. On the salvaged panels the architect Hesire was depicted with various objects and figures that had symbolic significance at the time (Fig. 1.20). Figure 1.20. Hesire panels For a long time, the purpose of the Hesire panels remained vague. At first, the Egyptologists considered these panels to be merely false doors. However, in the mid20th century, the situation surrounding the panels began to clarify. In the early sixties of the 20th century the Russian architect Shevelev ob Chapter 1 The Golden Section 33 served that the sticks the architect holds in his hands on one of the panels relate between themselves as 1 : 5 , that is, as the ratio of the small side of the 1:2 rectangle (“double square”) to its diagonal. Basing on this observation, another Russian architect Shmelev made a very precise geometrical analysis of the Hesire panels. This analysis led him to sensational discovery [36]. Having explored the Hesire panels, Igor Shmelev came to the conclusion: “Now, after the comprehensive and justified analysis, by using the proportion method, we have good ground to assert that the Hesire panels are the harmony rules encoded in geometric language… So, in our hands we have the concrete evidence, which shows us by “plain text” the highest standard of creative thought of the Ancient Egyptian intellectuals. The artist who created the panels with amazing accuracy, exquisite refinement and masterly ingenuity demonstrated the rule of the Golden Section in its broadest range of variations. It gave birth to the “GOLDEN SYMPHONY” presented by the ensemble of highly artistic works, which testifies not only the ingenious talents of their creator, but also convincingly proves that the author was aware of the secret of harmony. This genius was the “Golden Affair Craftsman” by the name Hesire.” But who was Hesire? Ancient texts inform us that Hesire was “a Chief of Destius and a Chief of Boot, a Chief of Doctors, a Scribe of Pharaoh, a Priest of Gor, a Main Architect of Pharaoh, a Supreme Chief of South Tens, and a Carver.” Analyzing the regalia of Hesire, Shmelev paid particular attention to the fact that Hesire was the Priest of Gor. Gor was considered by the Ancient Egyptians as the God of Harmony, and therefore, to be the Priest of Gor he had to execute the functions of the keeper of Harmony. As it follows from his name, Hesire was elevated to the rank of God of Ra (the God of the Sun). Shmelev assumes that Hesire was elevated to this rank rewarded for his “development of aesthetic … principles in the canon system reflecting the harmonic fundamentals of the Universe … The orientation to the harmonic principle discovered by the Ancient Egyptian civilization was the way to the unprecedented flowering of culture; this flowering falls into the period of Zoser, pharaoh when the system of written signs was complete ly implemented. Therefore it is possible to assume that Zoser’s pyramid be came the first experimental pyramid, which was followed by construction of the unified complex of the Great pyramids in Giza according to the program designed under Hesire’s supervision” [36]. Thus, Shmelev states: “It is only necessary to recognize that the Ancient Egypt civilization is the supercivilization explored by us extremely superfi cially and it demands a qualitatively new approach to the development of its richest heritage… The outcomes of researches of Hesire’s panels demonstrate Alexey Stakhov THE MATHEMATICS OF HARMONY 34 that the sources of modern science and culture are in boundless layers of history feeding creativity of the craftsmen of our days by great ideas, which for long time inspired the aspirations of the outstanding representatives of the mankind. And our purpose is to keep a unity of a connecting thread” [36]. 1.8.2. The Mysteries of the Egyptian Pyramids This infinite neverending uniform sea of sand, with infrequent dried bushes, and the barely visible tracks from a camel covered by sand, is a wasteland under an incandescent sun. Everything around seems dull and covered by fine sand. Suddenly, as if beset by a mirage, to our amazement we see the pyramids (Fig. 1.21) whose fancy stone figures are directed toward the Sun. They as tonish our imagination by their vast size and perfect geometric forms. A Pharaoh’s authority in Ancient Egypt was tremendous; divine honors were given to him, and a Pharaoh was called the “Great God.” The GodPha raoh was a progenitor of the country and a judge of people’s fate. The cult of the dead pharaoh had great significance in the Egyptian religion. The monu mental pyramids (in addition to their more profound philosophical and math ematical significance) were constructed for the preservation of a pharaoh’s body, spirit, and for extolling his authority. Constructed by human hands, they deserve to be included amongst the Seven Wonders of the World. The pyramids had multiple functions. They served not only as vaults of a Pharaoh’s mortal remains, but they also were the attributes of his majesty, power, riches of country, the monuments of culture, the country’s history, and items of information about the pharaoh, his people, and life. It is clear that the pyramids held deep “scientific knowledge” embodied in their forms, sizes, and orientation of terrain. Each part of a pyramid and each element of its form were selected carefully to demonstrate the high level of knowledge of the pyramid creators. They were constructed for mil lennia, “for all time,” and for this reason, the Arabian proverb says: “All in the World are afraid of a Time. However, a Time is afraid of the Pyramids.” Among the gigantic Egyptian pyramids, the Great Pyramid of the Pharaoh Cheops (Khufu) is of special interest (Fig. 1.21). Before the beginning of the analysis of the form and dimensions of Cheops’ Fig. 1.21. Cheops’ Pyramid in Giza Pyramid it is necessary to introduce the Chapter 1 The Golden Section 35 Egyptian measurement system. The ancient Egyptians used three units of mea surement: the “elbow” (466 mm) which equaled 7 “palms” (66.5 mm), which, in turn, was equal to 4 “fingers” (16.6 mm). Let us analyze the dimensions of Cheops’ Pyramid (Fig. 1.22) employing considerations given in the remarkable book of Ukrainian scientist Nikolay Vasutinsky [31]. The majority of researchers believe that the length of the side of a pyra mid’s base is, for example, GF is equal to L = 233.16 m. This value corresponds almost precisely to 500 “elbows.” The full accordance to the 500 “elbows” will be, if we take the length of the “elbow” equal to 0.4663 m. The altitude of the Pyramid (H) was estimated by researchers variously from 146.6 m up to 148.2 m and depending on the adopted pyramid’s altitude, all ratios of its geometric construction will change considerably. What is a cause for the dif ferences in the estimation of a pyramid’s altitude? Strictly speaking, Cheops’ pyramid is truncated. Today, its apex is approximately 10 × 10, but 1 century ago, it was 6 × 6. Apparently, the top of the pyramid was dismantled. In estimating a pyramid’s height, it is necessary to consider physical fac tors, such as “shrinkage” of its construction. Under the enormous pressure (reaching 500 tons on 1 m2 of the ground) exerted by its weight, the height of most pyramids would have decreased in comparison to its initial designed val ue. What was the initial height of the Pyramid? This height can be reconstruct ed if we can discover the main or key “geometrical idea” of the Pyramid. In 1837, English Colonel Howard Vyse measured the inclination angle of the Pyramid faces: it appeared to be α = 51°51'. A majority of researchers recognize this value today. The indicated value of the inclination angle corre sponds to the tangent equal to 1.27306. This value corresponds to the ratio of the pyramid’s altitude AC, which is half of its base СВ (Fig. 2.22), that is, A AC / CB = H / ( L / 2) = 2H / L . Researchers were in a big sur prise! If we consider the square H root of the golden mean, that is, E τ , we obtain the following out come τ = 1.272. Comparing this value with the value of 51°50' 51°50' D B tgα = 1.27306, we see that C these values are very close. If we accept the angle α = 51°50', that is, we de G F crease it by one arc minute, Figure 1.22. A geometric model of Cheops’ pyramid the value of tgα will become Alexey Stakhov THE MATHEMATICS OF HARMONY 36 equal to 1.272, that is, it will be exactly equal to the value of τ . It is necessary to note that in 1840, Colonel Vyse repeated his measurements and corrected the value of the angle to α = 51°50'. These measurements led researchers to the following hypothesis: the ratio AC/CB = 1.272 underlies the triangle АСВ of Cheops’ Pyra mid! This ratio is characteristic for the above golden right triangle (Fig. 1.11). Then, if we accept the hypothesis that the golden right triangle is the main geo metrical idea of Cheops’ Pyramid, it is then possible to calculate the initial height of Cheops’ Pyramid. It is equal to H = ( L / 2) × τ = 148.28 m . Now, we deduce other relationships for Cheops’ Pyramid following from this golden hypothesis. In particular, let us find the ratio of the external area of the Pyramid to its base. For this purpose, we accept the length of the leg СВ for the unit, that is, CB = 1, however, the length of the side of the Pyramid base is GF = 2, and the area of the base EFGH will be equal to SEFGH = 4. Now, let us calculate the area of the lateral side of Cheops Pyramid. As the altitude AB of the triangle AEF is equal to τ, then the area of each lateral side will be equal to SD = τ , and the common area of all four lateral sides of the pyramid will be equal to 4τ, while the ratio of the external area of the pyramid to its base will be equal to the golden mean. This is the main geometrical secret of Cheops Pyramid! Analyses of other pyramids confirm that the Egyptians always intended to embody their most significant mathematical knowledge in the pyramids they built. In this respect, the Pyramid of Chephren (Khafre) is also rather interest ing. The measurements of Chephren’s Pyramid show that the inclination angle of the lateral sides is equal to 53°12', which corresponds to leg’s ratio of the right triangle: 3:4. This leg’s ratio corresponds to the wellknown right triangle with the side ratio: 3:4:5, this triangle is called the “perfect,” “sacred,” or “Egyp tian” triangle. According to testimony from historians, the “Egyptian” triangle had a magical sense. Plutarch wrote that the Egyptians compared the universe to the “sacred” triangle; they symbolically ascribed the vertical leg to the hus band, the base to the wife, and the hypotenuse to the child born from them. According to the Pythagorean Theorem for the triangle 3:4:5 we have: 2 3 +42 = 52. Possibly, this famous theorem was perpetuated in Chephren’s pyr amid based on the 3:4:5 triangle. It is difficult to find a more appropriate ex ample to demonstrate The Pythagorean Theorem — which evidently was wellknown to the Egyptians long before its rediscovery by Pythagoras. Thus, the ingenious architects of the Egyptian Pyramids sought to astonish future generations and fulfilled this goal by selecting the golden right triangle as the “main geometrical idea” of Cheops’ Pyramid, and the “sacred” or “Egyp tian” right triangle as the “main geometrical idea” of Chephren’s Pyramid. Chapter 1 The Golden Section 37 1.9. The Golden Section in Greek Culture 1.9.1. Pythagoras Pythagoras is one of the most celebrated people in the history of science. His name is known to each person who has studied geometry and been introduced to the Pythagorean Theorem. The famous philosopher, scientist, religious and ethical reformer, influential politician, “demigod” — in the eyes of his followers, Phythag oras personifies ancient wisdom. Coins with his image, which were made in 430 420 B.C., testify to the exclusive popularity of Pythagoras. In the 5th century B.C., this was unprecedented! Pythagoras was the first, among the Greek philoso phers, to have a special book dedicated to him. His scientific school is internationally known. He had organized it in Croton, a Greek colony in northern Italy. The “Pythagorean School,” or “Pythagorean Union,” was simultaneously a philosophical school, a political party, and a religious brotherhood. Entrance to the “Pythagorean Union” was very demanding. Each, who entered the “Pythagorean Union,” would be re fused personal property for the benefit of the Union; they undertook to not spill blood, not eat meat, and Pythagoras protect the secret of their doctrine. The members of (Born c.569 B.C., died c.475 B.C.) the Union were prohibited to train others for money. The Pythagorean doctrine focused upon harmony, geometry, number the ory, astronomy, and other topics, but the Pythagoreans, most of all, appreciat ed the results that were obtained in the theory of harmony because it con firmed their idea that “numbers determine everything.” Some ancient scien tists assume that the concept of the golden section was borrowed by Pythag oras from the Babylonians. Many great mathematical discoveries were attributed to Pythagoras — some perhaps undeservedly. For example, the famous geometric “theorem of squares” (Pythagorean Theorem), which sets a ratio between the sides of a right triangle, was known to the Egyptians, the Babylonians, and the Chinese long before Pythagoras. However, discovery of “incommensurable line segments” is considered to be the main mathematical discovery of the Pythagoreans. In investigating the ratio of a diagonal to the side of a given square, the Pythagoreans proved that Alexey Stakhov THE MATHEMATICS OF HARMONY 38 its ratio cannot be expressed as the ratio of two natural numbers, that is, this ratio is “irrational.” This discovery caused the first crisis in the history of mathematics, as the Pythagorean doctrine about integervalued basis of all existing things could no longer be accepted as true. Therefore, the Pythagoreans attempted to keep the discovery a secret; they created a legend about the death of one of the Pythagoreans — who divulged the secret of this discovery. Also attributed to Pythagoras were a number of the other geometrical dis coveries, namely, the theorem about the sum of the interior angles of a triangle, and the problem about segmentation of the plane into regular polygons (trian gles, squares and hexagons). There is a view that Pythagoras discovered some or all of the primary regular “spatial” figures, that is, the five regular polyhedrons. Why was Pythagoras so popular already during his life? The answer to this question follows from interesting facts from his biography. According to legend, Pythagoras went to Egypt and lived there for 22 years to learn from the wisdom of the priestscientists. After studying all Egyptian sciences, in cluding mathematics, he moved to Babylon, where he lived for 12 years and was introduced to the scientific knowledge of Babylonian priests. Legends also attribute Pythagoras with a visit to India. It is possible because Ionia and Indiaat that timehad business relations. He returned home about 530 B.C. Pythagoras attempted to organize his philosophical school. However, due to undetermined reasons, he abandoned Samos and settled in Crotona Greek colony in North Italy. Here, Pythagoras organized his school. Thus, the outstanding role of Pythagoras in the development of Greek science consists in fulfillment of the historical mission of knowledge transmis sion from the Egyptian and Babylonian priests to the culture of ancient Greece. Thanks to Pythagoras, who was, without any doubt, one of the leading think ers of his time, the Greek science could obtain new knowledge in the fields of philosophy, mathematics, and natural sciences. This new scientific knowledge contributed to the rapid development and augmentation of Greek science in ancient Greek culture. In developing the idea about the role of Pythagoras in the historical de velopment of the Greek science, Igor Shmelev, in the brochure Phenomenon of Ancient Egypt [36], wrote: “His world renowned name the Croton teacher obtained after the rite of “consecration.” This name is compounded of two halves and it means “The Prophet of Harmony” because the “Pythians” in Ancient Greece were pa gan priests who predicted a future and Gor in Ancient Egypt personified harmony. So in the last years of their civilization, the Egyptian priests trans mitted their secret knowledge to the representatives of the new civiliza Chapter 1 The Golden Section 39 tion, symbolically cementing in one person the Union of man’s and woman’s origins, the bastion of Harmony.” It is known that the Pentagram has always attracted special attention from the Pythagoreans; they considered it the symbol of health. There is a legend surrounding the pentagram: when a guest in a foreign land, one of the Pythagore ans lay on his deathbed, unable to pay the man helping him. He asked the man to draw the pentagram on his dwelling, hoping that this symbol would be seen by one of the Pythagoreans. Some years later, one of the Pythagoreans saw this symbol and the kind host received a rich compensation. Already, the fact that the Pythagoreans chose the “pentagram,” brimming with golden sections, as the main symbol of their Fraternity, is another confir mation that the Pythagoreans knew and esteemed the golden section. The Russian researcher Alexander Voloshinov wrote [137]: “The Pythagoreans gave special attention to the pentagram, the fivepoint ed star that is formed by diagonals of the regular pentagon. In the pentagram, the Pythagoreans found all proportions wellknown in antiquity: arithmetic, geometric, harmonic, and also the wellknown golden proportion, or the golden ratio. A perfection of mathematical forms of the pentagram finds a reflection in the perfection of its form itself. The pentagram is proportional and, hence, beau tiful. Probably owing to the perfect form and the wealth of mathematical forms, the pentagram was chosen by the Pythagoreans as their secret symbol and a symbol of health. Thanks to the Pythagoreans, the fivepointed star is today a symbol for many states and is on the flags of many countries of the world.” 1.9.2. The Idea of Harmony in Greek Culture The idea of harmony based on the golden mean became one of the most fruitful ideas in Greek art. Nature, taken in a broad sense, includes the human creative patterns of art, where the same laws of rhythm and harmony domi nate. Aristotle states: “Nature aims for the contrasts and from them, instead of similar things, Nature forms a consonance …. This combines a male with a female and thus the first public connection is formed through the connection of contrasts, instead of similarities. Also in art, apparently, by imitating Nature the artist acts in the same way. Namely the painter makes the pictures conform to the originals of Nature by mixing white, black, yellow and red paints. Music creates the unique harmony by mixing different voices, high and low, long and short, in congregational singing. Grammar creates its entire art from the mixture of vowels and consonants.” Alexey Stakhov THE MATHEMATICS OF HARMONY 40 Major Minor To take the subject and eliminate all that is superfluous is the plan of the ancient Greek artist. This is the main idea of Greek art, where the golden section lay at the center of the canon of aesthetics. The theory of proportion is the foundation of Greek art, and, the problem of proportionality could not have escaped the thought of Pythagoras. Among the Greek philosophers, Pythagoras was the first who attempted mathematically to understand the essence of musical or harmonic ratios or intervals. Pythagoras knew that the intervals of the octave can be expressed by natural numbers, which fit the corresponding oscillations of the string, and these numerical relations were put forth by Pythagoras as the basis of his theory of musical harmony. Pythag oras was credited with possessing knowledge of arithmetic, geometric, har monic proportions, and the law of the golden section. Pythagoras paid special attention to the golden section by choosing the pentagram as the distinctive symbol of the “Pythagorean Union.” By developing further the Pythagorean doctrine about harmony, Plato analyzed the five regular polyhedrons (the so called Platonic solids) and emphasized their ideal beauty. As to the main requirements of beauty, Aristotle proposed order, proportion ality, and size limitation. The order arises when certain relations and proportions hold between the whole and its parts. In music, Aristotle recognized the octave as the most beautiful consonance; he considered that the ratio of oscillations be tween the basic tone and its octave is expressed by the first natural numbers: 1:2. He also maintained that in poetry, the rhythmic relations of a verse are based on small numerical relations, and by this, a beautiful impression is expressed. Except for the simplicity based upon the commensurability of separate parts and the whole, Aristotle, as well as his mentor Plato, recognized that the highest beauty of perfect figures and propor tions was based upon the golden section. The ancient Harmony theory, based on the prin ciple of the division in extreme and mean ratio (the golden section), became the “launching pad,” upon which, subsequently, a harmony concept was devel oped in the science and art of European culture. 1.9.3. The Golden Section in Greek Sculpture For a long time, the creations by the great Greek sculptors Phidias, Polyclitus, Myron, and Praxiteles were considered to be the standard of beauty and har monious construction based on the human body. The Figure 1.23. Doryphorus by Polyclitus Chapter 1 The Golden Section 41 0,618 0,382 statue of Doryphorus, created by Polyclitus in the 5th century B.C., serves as one of the greatest achievements of classical Greek Art (Fig. 1.23). This statue is considered to be the best example for analysis of the proportions of the ideal human body established by ancient Greek sculptors. It is especially important because the name “Canon” was ascribed to this famous sculpture. The harmon ic analysis of the Doryphorus as given in the book Proportionality in Architec ture (1933) [10], written by the Russian architect G.D. Grimm, indicates the following connections of the famous statue to the golden mean : 1. The first golden cut division of the Doryphorus figure with its overall height M 0=1, relating segments M 1 and M 2, is at the navel. 2. The second division of the lower part of the torso, relating M 2 and M 3, passes through the line of his knee. 3. The third division, relating M 3 and M 4, passes through the line of his neck. The Venus de Milo (Fig. 1.24), a statue of the goddess Aphrodite, is one of the bestknown monuments of Greek sculptural art. This statue was created by Agesandr in the 2nd century B.C. The Goddess Aphrodite is represented halfnaked, so that her clothing — which wrap up her legs and the bottom of her torso, as if it were a pedestal for the open hands, which are showing in movement (Fig. 1.24). During the Hellenic epoch, Aphrodite was one of the most favorite goddesses. From the Island Melos, Aphrodite is strict and restrained. It is thought that she stood on the high pedestal and looked at over the spectators. The Ve nus de Milo is a pearl of the Louvre; it is a standard of Figure 1.24. Venus de Milo female beauty in ancient Greece. 1.10. The Golden Section in Renaissance Art 1.10.1. The Idea of the “Divine Harmony” in the Renaissance Epoch The idea of Harmony belonged to such ancient conceptual ideas, which attributed to the interest of the church. According to Christian doctrine, the universe was created by God and submitted to its will unconditionally. The Christian God, at the creation of the universe, was guided by mathe Alexey Stakhov THE MATHEMATICS OF HARMONY 42 matical principles. In Renaissance culture, Catholic doctrine included the math ematical plan to which God created the universe. In the opinion of the American historian of mathematics, Morris Klein, the close merge of the reli gious doctrine about God as the cre ator of the universe, and the antique idea of numerical harmony of the uni verse, became one of the major sta ples of the Renaissance culture [6]. Mainly, the objective of Renaissance science is presented in the following text from the wellknown astrono mer Jogannes Kepler: “The overall objective of all re Figure 1.25. Holy Family by Michelangelo searches of the external world should be to discover the rational order and harmony, which was embodied by the God in the Universe and then was presented for us by the God on mathematics language.” The art of the Renaissance period (espe cially paintings) is substantially connected to topics from the Bible. The picture Holy Fam ily by Michelangelo is a bright example of such topical art. This picture is a fairly recog nized masterpiece of WestEuropean art. The picture is often named “Tondo Doni” because, firstly, the picture belonged to the Doni fam ily in Florence, and secondly, it had a round form (in English “tondo”). After the harmon ic analysis of this picture, researchers found that a compositional construction of the pic ture is based on the pentacle (Fig. 1.25). The picture Crucifixion, by Rafael San ti (14831520), is another example of a pic ture based on the topics of the Bible. The harmonic analysis of this picture (Fig. 1.26) showed that the compositional plan of the picture is based on the golden isosceles tri Figure 1.26. Crucifixion by Rafael Santi angle (Fig. 1.18). Chapter 1 The Golden Section 43 1.10.2. “Vitruvian Man” by Leonardo da Vinci During the age of the Renaissance, attempts to create an ideal model of a harmoniously developed human body were continuing. It is known, that a sweep of human hands approximately is equal to a human growth. It means that the human fig ure can be inserted into a square and a circle. The ideal human figures, created by Leonar do da Vinci and Durer, are widely known. For a long time, the opinion existed that the “pen tagonal” or “fivefold” symmetry, which is characteristic for flora and animals, is shown in the structure of the human body. The hu man body can be considered as a sample of the pentagram, where the human head, two hands, and two legs are, as if, beams of the pentagonal star. Such model found a reflec tion in the constructions of Leonardo da Vin ci and Durer, in particular, in the figure of the wellknown Vitruvian Man by Leonardo da Figure 1.27. Vitruvian Man Vinci (Fig. 1.27). by Leonardo da Vinci 1.10.3. “Mona Lisa” by Leonardo da Vinci The Renaissance, in the cultural history of the Western and Central European countries, is a transitive epoch, whereby it progressed from the Medieval culture to the culture of a new era. A humanistic world outlook and a return toward the antique cultural heritage, as though, “Re naissance” of the ancient culture, are the most typical features of this epoch. The Renaissance is characterized by large scientific shifts in the field of natural sciences. The close connection to art is a specific feature of science of this epoch, and this feature was sometimes expressed by one creative person. Leonardo da Vinci, the outstanding artist, scientist, and engineer of the Renaissance, was a brilliant example of a manysided per son. By his nature, Leonardo da Vinci had enviable health, was goodlook ing man, tall, and blueeyed. Leonardo was born on the 15th of April under the Star of Mars, and possibly, therefore, he possessed a huge force and man’s valour. He sang marvelously by composing melodies and verses for his listeners. He played many different musical instruments; moreover, he created new musical instruments. Alexey Stakhov THE MATHEMATICS OF HARMONY 44 But art and science were the main spheres of Leonardo’s creative work — where Leonardo showed his genius talent. For Leonardo’s art works, his contemporaries and descen dants gave such definitions, as “genius,” “divine,” “great,” but the same words can be used to describe Leonardo’s sci entific discoveries. He invented a military tank, a helicopter, an underwater ship, a parachute, an automatic weapon, a diving helmet, an elevator, and so on. He solved the most Leonardo da Vinci complicated problems of acoustics, botany, and cosmology; he invented an hour pendulum one century earlier than Galileo, and he developed a mechanics theory — among other things. In his relationship with Universe Harmony, Leonardo expressed the following words: “All earth, mountains, woods, and seas form a whole, in which each thing feeds the other one, all things are cooperated and interconnected, supported, but at the same time destroyed and updated.” Seemingly, each visitor of the Louvre in Paris tries to find the famous “Mona Lisa” (“Jokonda”) by Leonardo da Vinci. The Great Artist drew this portrait tense ly and long. He made many sketches; he paid attention to the pose of Mona Lisa, to the turn of her head, and to the position of her hands. Italian artist Giorgio Vazari (15111574) tells in his “Biographies,” that during the painting of Mona Lisa, Le onardo invited singers, musicians, and clowns to his studio to support the cheerful mood of the young woman and to have an opportunity to watch a changeable expression of her face. And only after four years of intense work he, at last, could present to cultural community his worldfamous “Mona Lisa” (“Jokonda”). It is considered, that a secret of Jokonda’s charm is in the variability of her smile. There is an opinion that the woman represented in the picture had lost a child, several months ago. The dark color of her clothes speaks about this. She sits in a quiet pose combining her hands in her lap, but her face is full of imperceptible movement: her lips tremble in light smile, and her smiling eyes attract spectators attentively and derisively. In those days, the lips, slightly opened in corners of the mouth, were considered as an attribute of elegance. Jokonda’s hardly appreciable smile, gentle and mysterious, lights up the picture. Mysteriousness of the image is strengthened by the background, a mountainous silveryblue landscape. What is the reason for Jokonda’s charm? The search for the answer to this question continues. By creating the masterpiece, Leonardo used a secret known for many portraitists: the vertical axis of the picture passes through the center of the left eye what should cause the feeling of excitation of the spectator, that is, in the picture the artist used a “Principle of Symmetry,” But, possibly, the reason is in another. The picture of the genius artist attracted attention of the researchers. After precise analysis, they found two golden triangles one with its base stand Chapter 1 The Golden Section 45 ing on the bottom of the picture, another with its base touching the top of the picture and the top touching the bottom of the picture, and their com mon height crosses the center of Jokonda’s left eye. (Fig. 1.28). Further harmonious analysis of the picture showed, that the center of Jokonda’s left eye is on the crossing of the two bisectors of the upper golden triangle, which, on the one hand, bisect the angles at the base of the upper golden triangle, and, on the other hand it divides the sides of this golden triangle in the golden section. Thus, Leonardo used in the picture not only the “Prin ciple of Symmetry,” but also the “Golden Section Principle.” What is considered the apex of Leonardo’s Figure 1.28. Harmonic Analysis creativity, this picture is thought as the crys of Mona Lisa (“Jokonda”) tallization of his genius, innermost thoughts, by Leonardo da Vinci and inspiration. Very little is known about Mona Lisa, except for several insignificant facts, therefore, it is difficult to answer very important questions often asked and discussed: whether she was simply a beautiful model for Leonardo or she was his muse and even his love. There are some facts, which confirm the last assumption, and this, probably, explains a special magic of the picture. But it is clear that Mona Lisa was that woman who inspired Leonardo to create this unique masterpiece inspiring thousands of people over many centuries. A huge number of legends are connected with this wellknown picture. We will begin from the history of its creation. Giorgio Vazari, in his “Biographies,” wrote the following: “Leonardo undertook to execute for Franchesco Del Jokon do the portrait of his wife Mona Lisa.” Some researches assume that Vasari, apparently, was mistaken. The newest investigations showed that in the pic ture another woman, not Mona Lisa, the wife of the Florentine nobleman Del Jokondo, was presented. Many researchers speculate why Del Jokondo refused the portrait of his wife, but in reality the portrait became the property of the artist, and this fact is an additional argument favoring the fact that Leonardo represented another woman and not Mona Lisa. However, there is another legend why Mona Lisa’s portrait was found at Leonardo — simply Leonardo had drawn two copies of this famous picture. Probably, the creation of this picture is connected with some secret in Leonardo’s life. Leonardo’s riddle begins with his birth. As is known, Leonar Alexey Stakhov THE MATHEMATICS OF HARMONY 46 do was an illegally born son of a woman, about which almost nothing is known. It is known only that her name was Katerina and that she was the owner of a tavern. Not much more information is known about Leonardo’s father. Mr. Pero, Leonardo’s father, was 25 years old when Leonardo was born. He was the notary, and he possessed impressing male merits: he had lived till the age of 77s, had four wives (three of them he buried), fathered 12 children and the last child was born when he was 75 years old. In the Renaissance, a relationship with illegitimate children was tolerat ed. Leonardo, at once, was recognized by the family of his father. However, in the house of his father he was taken away after awhile. Soon after his birth, Leonardo was sent together with Katerina to the village Anhiano located near to the city of Vinci. He remained there about four years. During these years, Mr. Pero married his first wife — a 16yearold girl — who had a higher social position than Leonardo’s mother. The young wife had turned out fruitless. Probably, for this reason, Le onardo at the age of almost 5 was taken in his father’s city house, where at once he was cared by his numerous relatives: grandfather, grandmother, fa ther, uncle, and foster mother. During his life, Leonardo always remembered his native mother Katerine who surprisingly resembled Mona Lisa, the wife of the Florentine merchant Jokondo, and, probably, this fact became the main reason of Leonardo’s desire to create Mona Lisa’s portrait. He embodied into this picture everything that was cheerful, light, and clear to him. He embodied all of a son’s love to his mother Katerine in this wellknown picture, which defined the development of painting for many centuries forward, and we should thank God that Le onardo da Vinci met Mona Lisa during the final stage of his creative way. 1.11. De Divina Proportione by Luca Pacioli 1.11.1. Luca Pacioli The spiritual heritage of Greece, Rome, and Byzantium were marvelous ly combined in the new Renaissance lead by its Titans of gigantic intellect and artistic talent. “Titan” is the most appropriate term for the likes of Le onardo da Vinci, Michelangelo, Nicholas Copernicus, Albrecht Durer, Raphael, Bramante and many others. Luca Pacioli, the Italian mathematician of the Chapter 1 The Golden Section 47 Renaissance, collaborated with Leonardo da Vinci, was the sem inal contributor to the field now known as accounting, takes a place of pride amongst the Titans. He was also called Luca di Borgo after his birthplace, Borgo Santo Sepolcro, Tuscany, where he was born in 1445. Luca Pacioli is rightfully called “The Father of Account ing” and “The Unsung Hero of the Renaissance.” In fact, the Fransican Friar Luca Pacioli was one of the most re markable people of his epoch, but unfortunately was one of Luca Pacioli (1445 1514) the least wellknown. This is surprising because his work as a master of math is brilliant and revolutionary and continues to affect us. Although we know little of Pacioli’s early life, there is evidence that he began to study in the art studio of the artist Piero della Francesca. Amongst his contemporaries, Pierro della Francesca was known as a mathematician and geometer, as well as, an artist. Today he is chiefly appreciated for his art. Through Piero, Pacioli gained access to the Count of Urbino and the library of Federico where he gained access to thousands of books. This allowed Luca to expand and deepen his knowledge of mathematics. Piero also would later introduce Pacioli to his new mentor, Leon Battista Alberti, who became the second great man in Luca Pacioli’s life. The following words of Alberti fell deeply into Luca’s consciousness: “We may conclude Beauty to be such a Consent and Agreement of the Parts with the Whole in which it is found, as to Number, Finishing and Collocation, as Congruity, that is to say, the principal Law of Nature requires. This is what Architecture chiefly aims at, and by this she obtains her Beauty, Dignity and Value.” Alberti introduced Luca to Pope Paul II. Pope Paul encouraged Luca to dedicate his life to God and become a monk. When Alberti died in 1472, Luca took the vows of a Franciscan Minor. In 1475 Pacioli began to work at the University of Perugia, receiving a professorial chair in 1477. He remained at the University for six years while becoming the first lecturer holding a chair in math in this University. While there Pacioli wrote a mathematical manuscript dedicated to the “Youth of Perugia.” After he left Perugia, he took up more traveling and wandered throughout Italy, but in 1486 was called back to the University of Perugia by the Franciscans. It was at this time that he started calling himself “Magister,” or “Master” — the equivalent of what we today refer to as a full professorship. Luca wrote his famous work Summa de Arithmetica, Geometria, Propor tioni et Proportionalita, when he was 49 years of age. Luca did this out of the belief that mathematics was being poorly taught at this time. Alexey Stakhov THE MATHEMATICS OF HARMONY 48 One section of the book that was entitled Particularis de Computis et Scrip turis was dedicated to accounting. This book was referred to by some as “a catalyst that launched the past into the future.” Luca was the first person to explain the doubleentry system, or the Venetian method. This was absolute ly revolutionary and far ahead of its time, sealing for him the title “The Father of Accounting.” The Summa was recognized for many accomplishments and it became the most widely read mathematical work in all of Italy. After Leonardo Da Vinci read Pacioli’s Summa, he arranged for Luca to come to the Court of Duke Lodovico Maria Sforzo to tutor him in mathemat ical perspective and proportion. Luca joined Leonardo at the Sforza Court in 1496, whence began a sevenyear relationship that produced two enduring masterpieces. Under the direct influence of Leonardo da Vinci, Luca Pacioli began to write the book De Divina Proportione, pub lished in 1509. The book had a noticeable influence on his contemporaries and was one of the first fine exam ples of the Italian art of bookprinting. The historical significance of this book is that it was the first mathe matical book solely dedicated to the golden mean. Da Vinci used his artistic skills to illustrate Luca’s De Div ina Proportione, the second important Pacioli manu script. At the same time, Luca taught Leonardo per spective and proportionality. This knowledge would remain with him forever, and help him to create one of his greatest masterpieces, “The Last Supper.” Painted on the back wall of the dining hall at the Dominican Figure 1.29. convent of Santa Maria delle Grazie in Italy, it instant Luca Pacioli’s book ly became one of the most famous works of the Italian De Divina Proportione Renaissance (if not of all time). De Divina Proportione consists of three parts: the first part sets out the properties of the golden ratio, the second part is dedicated to the five regu lar polyhedrons, and the third part describes the applications of the Golden Section in architecture. Using Plato’s Republic, Timeaus, and Laws Pacioli sequentially infers the twelve different properties of the Golden Section. Pacioli characterizes them using such epithets as “exceptional,” “remark able,” “almost supernatural,” etc. Considering his view (and that of the an cients) that the Golden Section is the universal ratio that expresses a per fection of Beauty in Nature and Art, he names it the Divine Proportion and recognizes it as a “thinking tool”, the “aesthetic canon”, and the “Funda mental Principle of the Cosmos”. Chapter 1 The Golden Section 49 This book is one of the first mathematical works that attempted to provide a solid scientific basis fot the Christian doctrine about God, the creator of the Cosmos. Considering the properties of the golden mean inherent in God, Pacioli names it the “Divine Proportion.” 1.11.2. About Pacioli’s Plagiarism Roger HerzFischler wrote in his book [40]: “Perhaps no mathematician has plagiarized as much and with so little change in details and has stirred so much controversy and such vehement reactions as Luca Pacioli.” The main charge for plagiarism falls on Pacioli’s De Divina Proportione. Some have claimed that this book is a literal translation out of the Latin of Piero della Francesca’s manuscript De Quinque Corporibis. However, De Divina Proportione is not the only book of Pacioli that falls under this charge. According to Davis [40], Pacioli also derived dodecahedron and icosahedron problems from Piero’s Trattato and included them in his 1494 book, Summa de Arithmetica, Geometria, Proportioni et Proportionalita. It is now surprising why Pacioli’s books caused such vehement reaction. Euclid, for instance, could also be blamed for plagiarism in The Elements. Ac cording to the statement of the famous mathematician and historian Van der Waerden, a majority of the mathematical results stated in The Elements be longed to the Pythagoreans. The following quote from Roger HerzFischler’s book [40] is very interesting in regard to the aforementioned: “As a final comment on Pacioli’s “literary borrowings,” I mention Cardano’s 16th century comments … on Pacioli’s use in his Summa of a 1202 manuscript (Fibonacci?) dealing with algebra; and Agostini (1925) who, after discussing Pacioli’s word for word inclusion in his Summa of a 1481 mercantile text, says that one should not accuse Pacioli of plagiarism for having used in his book what is available to everybody and is not personal intellectual property.” Thus, following the logic of Pacioli’s accusers, all modern books on geom etry are plagiarized, as they could not be written without using Euclid’s Ele ments! Unlike artistic creative work, a distinctive feature of scientific, in par ticular, mathematical creative work, is that ideas of preceding fundamental scientific results can be used by all subsequent authors in new articles and books. Therefore, it is our responsibility as scientists, who are often in the same position, to stand up for the great Italian mathematician, Luca Pacioli, against such accusations of plagiarism. As mentioned earlier, Piero della Francesca was the first teacher and very good friend of Luca Pacioli. It is more Francesca’s mathematical works than his art works that influenced the Alexey Stakhov THE MATHEMATICS OF HARMONY 50 young Luca. There is absolutely nothing extraordinary in Luca’s use of Piero della Francesca’s mathematical ideas. At the very moment Pacioli began working on his book, De Divina Proportione; he was already a wellknown mathematician and the author of the book Summa de Arithmetica, Geometria, Proportioni et Proportionalita. The latter could well be considered to be the mathematical encyclopedia of the Renaissance. Considering these facts, it is hard to believe that the book De Divina Proportione is literal translation of Francesca’s manuscript De Quinque Corporibis. More likely, Luca Pacioli used it as the basis of his “second great book” De Divina Proportione. Leonardo da Vinci unquestionably was the illustrator of this remarkable book with its 60 geometrical figures. Perhaps somebody will argue that he borrowed them from Piero della Francesca’s works. Following the logic of the Pacioli accus ers, all future books in the field of the “Golden Section,” including the present book, should be recognized as a “plagiarism” of Pacioli’s book De Divina Pro portione, the first book focusing on the Golden Section in history! 1.11.3. Death and Oblivion of Luca Pacioli Already tired and ill, in 1510, Luca Pacioli was 65 years old. In the fore word of his unpublished book About the Forces and Quantities he wrote the sad phrase: “The last days of my life approach.” This manuscript is stored in the library of the Bologna University. Pacioli died in 1515 and is buried in the cemetery of his native town Borgo SanSepolcoro. After Pacioli’s death, his works fell into oblivion for almost four centuries. At the end of the 19th century, his works became internationally wellknown. After the 370year hiatus of recognition, grateful descendants in 1878 placed a memorial stone in Pacioli’s house of birth in Borgo San Sepolcro with the following inscription that reflects the essential significance of Pacioli: “For Luca Pacioli, who had da Vinci and Alberti as friends and advisers, who turned algebra into a science, and applied it to geometry, who lectured in doubleenrty bookkeeping, whose work was the basis and the norm for later mathematical research; for this great fellow citizen, the people of San Sepul cro, ashamed of their 370 years of silence, have placed this stone, in 1878.” The names of three geniuses of the Renaissance are mentioned in this inscription on the stone: Leonardo da Vinci, Leon Battista Alberti and Luca Pacioli. Their works became a valuable if not priceless contribution to the development of the theory of Harmony and the Golden Section! Pacioli’s memorable work is not forgotten in 20th and 21st centuries. A beautiful Monument to Luca Pacioli was sculpted out of white Carrara Chapter 1 The Golden Section 51 marble and placed in Sansepolcro, It aly in honor of the anniversary of the 1494 publication of the Summa. In re cognition of Pacio li’s great works, a 500 lira post stamp was issued, and nu merous conferenc es were held around the world. Memorial stone in Borgo San Sepolcro, 1878 In honor of 500 anniversary of Summa, 1994 1.12. A Proportional Scheme of the Golden Section in Architecture The 1935 book Proportionality in Architecture [10] by Russian architect Professor Grimm is well known in the theory of architecture. The purpose of the book is set forth in its Introduction: “In view of the exceptional significance of the golden section as the pro portional division that establishes a constant connection between the whole and its parts and gives a constant ratio between them, unachievable by any other division, it is foremost the normative basis upon which we will subse quently check the proportionality of both historical and modern monuments and constructions…. Alexey Stakhov THE MATHEMATICS OF HARMONY 52 Taking into consideration this general importance of the golden section in all manifestations of architectural thought, it is necessary to recognize the impor tance of it, as the basis of architectural proportionality theory in general.” Grimm illustrates the golden section as a division of line segment AB at point C into two unequal parts where the larger part CB is called major, and the smaller part AC is called minor. The whole line AB is to the major (segment CB) as the major (segment CB) is to the minor (segment AC), both in the golden ratio. Fol lowing a detailed analysis of the properties of the golden section, and in a manner not unlike Pacioli’s comparison of it to the qualities Divine, Grimm places the golden mean in the forefront of all other proportions. He writes: “In general it is necessary to recognize the extremely outstanding prop erty of the golden section, which cannot be reached by arithmetic mean pro portions and other divisions of the whole.” Grimm further illustrates examples of linear proportionality of the gold en division (see the Doryphorus statue, Fig. 1.23), through analysis of golden divisions of rectangles, triangles, circles, and golden spirals. And finally, he reviews volumetric proportionality through divisions of cubes, parallelepi peds, triangular prisms, and tetrahedral pyramids. “This analysis of the significance of the golden section and its exclusive properties for the theoretical solution of proportional division problems of linear, planar and volumetric masses, leads us to the following conclusion: For full proportional coordination of architectural structures representing a volu metric solution, it is necessary to have a golden proportioning, not only of its linear dimensions in vertical and horizontal, but also of the planar areas, and thence, of all volumes.” Grimm validates his theoretical researches in the field of the golden proportional scheme by architectural examples from classical art (Par thenon, Jupiter’s temple in Tunis), Byzantine monuments, and Italian Renaissance art and ar chitecture (Saint Peter’s Cathedral, Fig. 1.30; and the Calleoni monument). Grimm also analyzed structures of the Baroque period as they differ from the architecture of the Classical period and the Italian Renaissance, and at first sight may have been lacking the golden sec tion. He analyzed the Smolny Cathedral (Fig. 1.31) Figure 1.30. Saint Peter’s Basilica in Rome and its surrounding monastery in St. Petersburg, (architect Bramante) Russia, a shining example of Baroque style. The Chapter 1 The Golden Section 53 complex was commissioned in 1744 by Empress Eliza beth, a daughter of Peter the Great, and was construct ed in place of the Admiralty’s resin yard. Construction commenced in 1749 under the leadership of the archi tect Rastrelli, author of the Winter Palace, and the ba sic structure was completed in 1764. The magnificent bluewhite cathedral rising to a height of 93.7 meters stands at the center of the Smolny monastery com plex. The cathedral is stunning in its luxury, perfection of proportions, and variety of decorative forms. It pos sesses a strange, mystical peculiarity and magnificence Figure 1.31. Smolny that is strengthened as one approaches it. After ana Cathedral St.Petersburg, lyzing the cathedral, Grimm concluded that its archi Russia tecture was based upon the golden section. Although there is no generally accepted opinion amongst architects as to Grimm’s harmonious views, his editor’s foreword to Proportionality in Architecture states: “Nevertheless, his attempt at a general formulation of the golden section prin ciple as the basis of proportionality in a variety of architectural styles, support ed by analyses of material from ancient and European architecture, deserves to be published. All the more, Grimm’s book presents a historical sketch of the development and use of proportionality theory, and also a comprehensive math ematical statement of the principle of the golden section.” 1.13. The Golden Section in the Art of the 19th and 20th Centuries 1.13.1. Ivan Shishkin’s Picture “The Ship Grove” Ivan Shishkin (18321898) holds a place of honor in Russian painting. The entire history of Russian landscape painting in the second half of the 19th century is associated with his name. The art works of this eminent master took on special significance and became established as classics of Russian painting. Shishkin’s creative work is an exclusive phenomenon amongst the masters of painting. Like many Russian artists he possessed a tremendous native talent. Alexey Stakhov THE MATHEMATICS OF HARMONY 54 Major Minor Minor Major The picture The Ship Grove (Fig. 1.32), created by Shishkin in 1898, is his last work. It is considered a worthy cul mination to his original and creative ca reer. This picture has a perfect classical form, from the point of view of com pleteness and composition. The natural etudes created by Shishkin in his native land’s forests, where he found his own Figure 1.32. Harmonic analysis of idealistic synthesis of harmony and greatness, were thought to be the basis of his landscapes. In The Ship Grove the artist exemplified a thorough knowledge of Russian nature that was infused within him throughout his almost fifty years of creative life. It is obvious that the golden section law can be found in this wellknown picture, The Ship Grove (Fig. 1.32). For example, the brightly sunlit foreground pine divides the horizontal plane of the picture in the golden section. Further more, from the right of the pine there is the sunlit hillock. This hillock divides the vertical plane of the picture in the golden section. From the left of the foreground pine we can see many other pines. A division of the left part of the picture’s horizontal plane in the golden section suggests itself. Presence of the bright ver tical and horizontal lines that divide the picture in the golden section produces a balance and calmness consistent with the artist’s intention. 1.13.2. “Modulor” by Le Corbusier Le Corbusier (18871965) is an exceptional archi tect and theorist. His buildings became a triumph of the new architectural aesthetics with their vitality and humanism. After his book To Architecture was published in 1923, Le Corbusier became one of the twentieth century’s leading architectural theorists. During the Second World War Le Corbusier creat Figure 1.33. Le Corbusier’s Modulor ed his Modulor, a system of new proportional relations. He used the proportions of the human body as a principle of architectural met rics and, as a basis of his Modulor, he took not only the average height of a man, but also man’s measurements while seated, including lengths of the hand and foot. The Modulor is not just a theory; it is a practical guide for the employment of human proportions in architecture. Einstein was one of the first great scientists who appreciated the Modulor for its true value. He recognized that Corbusier’s Chapter 1 The Golden Section 55 system had practical importance not only for architec ture, but for other kinds of human activity as well. Har monic analysis of the Modulor (Fig. 1.33) leaves no doubt of the presence of the golden section. 1.13.3. The Building of the United Nations Headquarters in New York Figure 1.34. The Building of the United Nations Headquarters in New York The construction of this world famous building (Fig. 1.34) is associated with the name of Oscar Ni emeyer (19071989), wellknown twentieth centu ry Brazilian architect. He was the main adviser at the construction of the United Nations headquar ters in New York. Its main architectural idea of this building consists of the three golden rectangles. 1.13.4. A Picture “Near to the Window” by Konstantin Vasiliev Konstantin Vasiliev (19421976) is a modern Russian artist who unfortu nately died quite young. He lived and worked near Kazan in the village Vasilevo. His popularity is owed to his mixture of traditional symbolism and modernism through the painting of folklore themes from ancient and modern Russian histo ry. He was first introduced to the golden section at the Kazan school of art. Having studied the compositional principles of the golden section used by an cient Greeks, Konstantin decided “to investigate harmony through algebra.” He began exploring the possibility of applying the laws of harmony to the whole picture area in order to achieve maximal artistic expression of the image. He was moving the different segments expressing the golden section within the picture. Lastly, using the golden spiral, Vasiliev defined for himself how the hu man eye perceives the subject of the picture. Thus he would find how the picture should be constructed, and where he should set the imaginary main point of the picture so that it would absorb all the plot lines. Vasiliev’s Near to the window (1975) is a fine example of construction using the golden mean in all aspects of the picture (Fig. 1.35). We can only guess what the art ist was trying to tell us. The main idea of Figure 1.35. Vasilev’s picture Near to the window (1975) this picture and all of its ramifications lies Alexey Stakhov THE MATHEMATICS OF HARMONY 56 in the face of the young woman glowing with cleanliness, dignity and quiet wis dom. The artist framed her face around the golden point of the picture that is located on the crossing of the two golden lines, horizontal and vertical. Therefore, this compositional decision is one of the reasons for the feeling of harmony of the picture that manifests within all primordial beginnings and the beauty of Russian women. 1.14. A Formula of Beauty 1.14.1. The Golden Ratio in the External Forms of a Person Countless artists, poets and sculptors have admired the beauty of the human body! The French sculptor Rodin proclaimed, “The naked body seems to me beautiful. For me it is a miracle where there can not be anything ugly.” A human is considered the highest creation of Nature. Therefore, a human body at all times is recognized as the most perfect and worthy object of sculp tural art. The problem of correct proportional representation of the human body was always one of the most important factors in art. The golden mean played a leading role in the art canons of Leonardo da Vinci, Durer and other great artists. According to these canons the golden section relates the whole body to its two unequal subdivisions cut at the waistline. This is the most simple realiza tion of the proportional division of the human body (Fig. 1.36a). In addition, the golden proportion is often used by designers of clothes (Fig. 1.36b). b) a) Figure 1.36. The golden mean in human body (a) and clothes design (b) Chapter 1 The Golden Section 57 1.14.2. The Beauty of a Woman’s Face Repeated attempts have been made to analyze a woman’s face using the golden mean and pentagram (Fig. 1.37). Figure 1.37. A harmonious analysis of a woman’s face Numerous researchers have come to the general conclusion that a woman’s face is beautiful due to the golden section relations exemplified in it. The face of a woman displays an array of emotions that are an integrated element of her beauty. It is proven that a woman’s face fits the proportions of the golden section most fully when she smiles. Any woman is perceived as more beautiful with a warm smile than with a rigid face filled with anger, arrogance and disregard. Ad mirers of great feminine beauty would be advised to make note of these golden section principles of aesthetics. 1.14.3. Nefertiti During excavations in 1912, a German archaeological expedition digging in the deserted city of Amarna discovered a ruined house and studio complex. The building was identified as a sculptor’s workshop. One of the items found had the name and job title of Thutmose, court sculptor of Egyptian Pharaoh Akhenaten. Among the many sculptures recovered was the famous head of Queen Nefertiti (in ancient Egyptian meaning the “beauty is coming”). This sculpture is Alexey Stakhov THE MATHEMATICS OF HARMONY 58 recognized as a symbol of feminine beauty. After the exposition of Nefertiti in the Berlin museum, her fame rose as if she were a modern movie star rather than an ancient Egyptian queen. With a charming head, long harmonic neck, and direct, but gently outlined nose, the Nefertiti sculpture caused increased interest in Egyptian art, including its deep mystical past of a cult of priests and esoteric wisdom. Possibly, our irra tional century has selected Nefertiti as a symbol of unusual feminine beauty, feeling and as yet unrealized affinity for the grandeurs of ancient Egyptian cul ture. Belarusian philosopher Edward Soroko tried to determine what ideas were used by Thutmose in the sculpture of Nefertiti [138]. He thinks that Thutmose’s logic was very clear and simple. In ancient Egypt, harmony was the prerogative of the Divine order that dominated the universe, and geometry was the main tool of its expression. The Queen played the role of Goddess. Hence, her image, which personified the wisdom of the world, must have been formed with geometrical perfection and irreproachable harmony, beauty and clarity. As a matter of fact, the main idea of ancient Egyptian aesthetic philosophy was to glorify the eternal, the measured, and the perfect in a constantly changing universe. Figure 1.38. Harmonic analysis of Nefertiti’s portrait Soroko made a harmonious analysis of Nefertiti’s sculpture and came to the conclusion that Thutmose had used the golden section principle as the basis of its design. During his analysis, Soroko found a harmonious system of regular geometric figures such as triangles, squares, and rhombi (see Fig. 1.38). Thus, he established that the parts of these figures, if put in order according to their sizes, are guided by one and the same ratio — the golden mean. The metric structure of Nefertiti’s statuette shows that there were an cient Egyptian sculptors consciously employing the principle of the golden section in their creative work. Soroko’s analysis is one of the more compelling confirmations of the role of the golden mean in ancient Egyptian art. Chapter 1 The Golden Section 59 1.15. Conclusion The geometrical problem of the “Division in Extreme and Mean Ratio” (DEMR) came to us from Euclid’s Elements. It was named later the golden sec tion, golden mean, golden number, golden proportion or divine proportion. DEMR appears throughout Euclid’s Elements and was used by Euclid for the geometric construction of the isosceles triangle with the angles 72°, 72° and 36° (the gold en triangle), regular pentagon and dodecahedron. This fact may have given rise to the widespread belief (see Proclus, commentator of Euclid’s Elements) that the main goal of the Elements was to describe the geometric construction and numerical interrelationship of the Platonic Solids. This means that the Pythagorean doctrine about the numerical Harmony of the Spheres was real ized in the greatest mathematical work of ancient science, Euclid’s Elements. As the history of science shows, the golden section appears diffused with in and throughout our cultural history. It became an aesthetic canon for the Egyptian, Greek and Renaissance cultures. During many millennia the gold en section was a subject of delight for the great scientists and thinkers includ ing Pythagoras, Plato, Euclid, Leonardo da Vinci, Luca Pacioli, Johannes Kepler, Allan Turing and countless others. A small list of some of the out standing works of art and masterpieces which are based on the golden section could contain Khufu’s Great Pyramid, the most famous of the Egyptian pyr amids, Nefertiti’s sculptural portrait, the majority of Greek sculptural monu ments, the magnificent Mona Lisa by Leonardo da Vinci, the works of Raphael, Shishkin and Konstantin Vasiliev, Chopin’s etudes, the musical works of Beethoven, Tchaikovsky and Bella Bartok, and the Моdulor of Corbusier. Alexey Stakhov THE MATHEMATICS OF HARMONY 60 Chapter 2 Fibonacci and Lucas Numbers 2.1. Who was Fibonacci? 2.1.1. Leonardo of Pisa Fibonacci Usually we associate the “Middle Ages” with the inquisition, forced conver sions, burning of witches and heretics, and bloody Crusades in search of “Christ’s Tomb.” Science was not in the spotlight. Under these circumstances, the appear ance of the 1202 mathematical book “Liber abaci” by Italian mathematician Le onardo of Pisa could not stay unnoticed by the scientific community. Who was the author of this book, and why were his mathematical works so important for Western European mathematics? To find the answers to these questions we have to return to his time and reproduce the historical epoch, in which Leonardo of Pisa, nicknamed Fibonacci, lived and worked. It is worth noting that the period between the 11th and 12th centuries was an epoch of the brilliant flowering of Arabian culture; however, this century was the beginning of its downfall. At the end of the 11th century, before the beginning of the Crusades, the Arabs were undoubtedly the most educated people in the world and went far beyond their Christian competitors. Arabian influence had pene trated to the West long before the Crusades. After the Cru Fibonacci sades, Arabian culture continued to influence the West and (c. 1170after 1228) this mingling of cultures began to erode the Arabian world. Western research ers were dazzled with the Arabian world’s art and scientific achievements. Their interest rapidly increased in Arabian geographical maps, algebra, astronomy tu torials, and architecture. During this time the great European mathematician of the Middle Ages, Leonardo of Pisa, nicknamed Fibonacci (son of Bonacci or son of the bull), lived and worked in Pisa. He admired Arabic science and math. Chapter 2 Fibonacci and Lucas Numbers 61 We know little about Fibonacci’s life. The exact date of his birth is un known, though he is thought to have been born around 1170. His father was a merchant and a government official who represented a new class of executives generated by the “Commercial Revolution.” At that time the city of Pisa was one of the largest commercial Italian centers that actively cooperated with the Islamic East. Fibonacci’s father traded in one of the trading posts founded by Italians on the northern coast of Africa. Thanks to that, he was able to give his son Fibonacci, the future mathematician, a good mathematical education in one of the Arabian educational institutions. Moritz Cantor, wellknown historian of mathematics, called Fibonacci “the brilliant meteor that flashed on the dark background of the Western Europe an Middle Ages.” Fibonacci wrote several mathematical works, including Liber Abaci, Liber Quadratorum, and Practica Geometriae. The book Liber Abaci is the best known of them. This book was printed twice during the life of Fibonacci, originally in 1202 and again in 1228. The book consisted of 15 sections covering the fol lowing topics: the new Hindu numerals and representation of numbers with their help (Section 1); multiplication, addition, subtraction and division of numbers (Sections 25); multiplication, addition, subtraction and division of fractions (Sections 67); finding the prices of the goods and their exchange, the rules of mutual aid and the rule of the “double false situation” (Sections 8 13); finding the quadratic and cubic roots (Section 14); and the rules related to geometry and algebra (Section 15). The book was intended as a manual for traders, though its significance far exceeded the bounds of trade practice. Fibonacci’s book is a sort of mathemati cal encyclopedia of the Middle Ages epoch. Section 12 is of particular interest. It makes up nearly one third of the book and, apparently, Fibonacci paid special attention to it and demonstrated great innovation within it. The bestknown Fibonacci problem is that of “rabbit reproduction.” Its solution purportedly re sulted in the discovery of the numerical sequence 1, 1, 2, 3, 5, 8, 13 ..., later called Fibonacci Numbers. This famous problem will be reviewed later. 2.1.2. Fibonacci and Abu Kamil Roger HerzFischler [40] pointed out that Fibonacci borrowed many mathe matical problems from the Arab mathematicians, in particular, Abu Kamil. How ever, in the summary of this comparison HerzFischler did not accuse Fibonacci of plagiarism. He wrote [40]: “Again when we turn to Fibonacci’s presentation of the problems from Abu Kamil’s, On the Pentagon and Decagon, we find new meth Alexey Stakhov THE MATHEMATICS OF HARMONY 62 ods of solution that again display a deep understanding and ability. We must con clude then that either there were several, now lost, works from which Fibonacci obtained his material or he was responsible for significantly raising the level of the applications of the properties of DEMR to various computational problems.” 2.1.3. Influence of Fibonacci’s Works on the Development of the European Mathematics Although Fibonacci was one of the great mathematical intellects in the histo ry of Western European mathematics, his contribution in mathematics continues to be understated. The Russian mathematician Professor Vasiliev, in his book In teger Number (1919), noted the merits of Fibonacci’s creative mathematical work: “The works of the educated merchant from Pisa were so far above and beyond the level of mathematical knowledge of the scientists of those times, that the influence of his work on mathematical literature only became recog nized two centuries after his death, at the end of the 15th century, when many of his theorems and problems were included by Luca Pacioli… and in the be ginning of the 16th century, when the group of talented Italian mathemati cians Ferro, Cardano, Tartalia, and Ferrari produced the beginnings of higher algebra thanks to the solution of the cubic and biquadrate equations.” It follows from this statement, that Fibonacci surpassed the Western Eu ropean mathematicians of his time by almost two centuries. His historical role for Western science is similar to the role of Pythagoras who acquired his sci entific knowledge from Egyptian and Babylonian sciences and then transferred them to Greek science. Fibonacci received his mathematical education in the Arabian educational institutions and transferred the Arabic knowledge of math to Western European science. Much of the knowledge acquired there, in par ticular, the ArabicHindu decimal notation, was introduced by him to West ern European mathematics. Thus, he provided the fundamentals for the fur ther development of Western European mathematics. 2.2. Fibonacci’s Rabbits 2.2.1. The “Rabbit Reproduction Problem” Fibonacci made a valuable contribution to science and the development of mathematics, however, the irony of his fate is that in modern mathematics he is Chapter 2 Fibonacci and Lucas Numbers 63 generally only known as the author of an unusual numerical sequence, the Fi bonacci numbers. He deduced that sequence when he was looking for the solu tion to the now famous rabbit reproduction problem. The formulation and solu tion of this problem is considered to be Fibonacci’s main contribution to the development of combinatorial analysis. In it Fibonacci anticipated the recur sive method that was recognized as one of the most powerful methods of solving combinatorial problems. The recursive relation achieved by Fibonacci is reput ed to be the first recursive relation in mathematical history. Fibonacci formu lated the core of the rabbit reproduction problem as follows: “A pair of rabbits were placed within an enclosure so as to determine how many pairs of rabbits will be born there in one year, it being assumed that every month a pair of rabbits produces another pair, pro vided that rabbits only begin to produce a new pair at their maturity, which is two Figure 2.1. Fibonacci’s rabbits months after their own birth.” For the solution to this problem (Fig. 2.1) we define a pair of the mature rabbits by A, and a pair of the newborn rabbits by B. We display the process of reproduction with two transitions that describe the rabbits’ monthly trans formations: A→AB (2.1) B→A (2.2) Note that the transition (2.1) simulates a monthly transformation of each pair of mature rabbits into two pairs, namely, the same pair of mature rabbits A and the newborn pair В. The transition (2.2) simulates the process of rabbit maturation when a newborn pair B is transformed into a mature pair А. Fur ther, by beginning with a mature pair A, the process of rabbit reproduction can be represented by Table 2.1. Table 2.1. Rabbit reproduction Date January, 1 February, 1 March, 1 April, 1 May, 1 June, 1 Pairs of rabbits A AB ABA ABAAB ABAABABA ABAABABAABAAB A 1 1 2 3 5 8 B 0 1 1 2 3 5 A+B 1 2 3 5 8 13 Note that in the columns A, B and A+B of Table 2.1 we can see the num bers, respectively, of mature (A), newborn (B) and total (A+B) rabbits. Alexey Stakhov THE MATHEMATICS OF HARMONY 64 Studying the A, B and (A+B) sequences, we find the following regularity: each number of the sequence is equal to the sum of the previous two numbers. If we now designate the nth number of the sequence that satisfies the rule as Fn, then the above general rule can be represented by the following mathematical formula: Fn=Fn1+Fn2. (2.3) Such formula is called recurrence or recursive relation. Note that the specific values of the numeric sequences generated by the recursive relation (2.3) depend on the initial values (the seeds) of the sequence F1 and F2. For example, we have F1 = F2 =1 for the A series and for this case the recursive formula (2.3) generates the following numerical sequence: 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, .... (2.4) For the B series we have: F1=0 and F2=1 then the corresponding numerical sequence for this case is as follows: 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, .... At last, for the (A+B) series we have: F1=1 and F2=2 then the correspond ing numerical sequence for this case is the following: 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, .... The numerical sequence (2.4) is usually called Fibonacci numbers. Fi bonacci numbers have a series of remarkable mathematical properties, which we will describe further. 2.2.2. About the Rabbits Why do rabbits occur in the history of mathematics? Rabbits are mammals of the hare family. The countries of Spain, France, and Italy are considered to be the native land of wild rabbits; from these countries the rabbits were introduced to other countries. Wild rabbits now inhabit the southern and middle parts of West ern Europe, and also Africa, Asia, Australia, New Zealand and America. A special feature of rabbits is their surprising rate of reproduction. Female rabbits become mature at the age of 34 months and are able to reproduce throughout the year. A female rabbit’s pregnancy lasts for 2832 days (on the average 30 days); this means, that, by considering process of rabbit reproduc tion, Fibonacci employed biological facts. However, mature rabbits may pro duce 8 to 10 newborn rabbits monthly. Thus, the rabbits reproduce more in tensively than Fibonacci suggested in his famous problem. This exclusive ability of rabbit reproduction can explain why many coun tries consider the “rabbit invasion” to be a “national tragedy.” Australia is one of Chapter 2 Fibonacci and Lucas Numbers 65 the examples. In 1837 one Australian farmer started a rabbit farm of 24 rabbits. The rabbits that reproduced and escaped to freedom could have nearly destroyed all the greenery of the continent. The Australian government succeeded in re ducing the number of rabbits by its resolute measures in the struggle with the socalled “longeared locust.” In Australia, war was declared against the rabbits and has lasted more than 150 years. Throughout the continent, Australians had constructed a kind of “Great Wall of China” of many hundreds of kilometers, an insuperable barrier for the rabbits. The war has progressed with variable suc cess. Wild Australian rabbits have learned to climb trees, becoming terrifyingly aggressive, capable of attacking the fields and kitchen gardens of the farmers. The “prolific tribe of rabbits,” which had influenced the famous Italian math ematician, now took the Italian island Ustina (to the North of Sicily) under siege. 100,000 rabbits overran the 1000 inhabitants of this small island. In con trast to the Australian inhabitants the native population of Ustina yielded to the rabbits without a fight; already one fifth of the inhabitants have emigrated from the island. It is necessary to remember the flip side of the “rabbit problem”; rabbit meat is considered useful and tasty. Italy, Fibonacci’s native land, is one of the largest producers of rabbit meat. Remembering this, the Italian (and nonItal ian) historians of mathematics should work to derive the answer to the fol lowing question: what was the main reason for Fibonacci introducing rabbit reproduction into mathematics: a love for mathematics or rabbit meat? 2.2.3. Dudeney’s Cows The English puzzlist Henry E. Dudeney (18571930) wrote several excel lent books on puzzles. In one of them he adapted Fibonacci’s rabbits to cows by making Fibonacci’s problem more realistic. He replaced months by years and rabbits with bulls (males) and cows (females). Dudeney states his cow problem as follows: If a cow produces its first shecalf at age of two and after that produces another single shecalf every year, how many shecalves are there after 12 years, assuming none die? In Dudeney’s opinion, this simplifies the problem and makes it quite realistic. 2.2.4. Honeybees and Family Trees As known, Fibonacci took a physical problem and simplified it in the way many mathematicians often do at first, namely, he simplified the problem and then observed to see what would happen. This series has many interesting Alexey Stakhov THE MATHEMATICS OF HARMONY 66 and useful applications, which we will see later. Now let us consider another reallife situation, honeybee reproduction, which can also be precisely mod eled on the Fibonacci series There are over 30,000 species of bees and most of them live solitary lives. The honeybees are the best known among them. They usually live in a colony called a hive and have an unusual “Family Tree.” In fact, there are many surpris ing features of the honeybees and, in this section, we will show how the Fi bonacci numbers can be used to construct the genealogical trees of honeybees. Before we discuss the Fibonacci series and honeybees, let us look at an unusu al fact about them. Not all honeybees have two parents! In a colony of honeybees there is one special bee called the queen bee. There are many workerbees, who are female too, but unlike the queen bee, they do not produce eggs. There are some male bees called dronebees. Dronebees are produced by the unfertilized eggs of the queen bee, so the male bees have only a mother, and no father. All female bees are produced when the queen bee has mated with a male bee and so fe male bees have two parents (Fig. 2.2). The females usually become workerbees. However, when some of them are fed with a special substance called royal jelly, each grows into a queen bee ready to go off to start a new colony. When this occurs, the Figure 2.2. Family tree of honeybees bees form a swarm and leave their hive in search of a place to build a new nest. Here we follow the convention of “Family Trees” where parents appear above their children, so the latest generations are at the bottom, and the higher up we go the older the generations. Such trees place all ancestors of the descendent at the top of the diagram. We would get quite a different tree if we listed all the descendants (offspring) of an ancestor in a way similar to the rabbit problem, where we placed all descendants of the initial pair at the bottom. Now, let us look at the family tree of the malebee. 1. The male bee has 1 parent, the queen bee. 2. The male bee has 2 grandparents because his mother (the queen bee) had two parents, a male and a female. 3. The male bee has 3 greatgrandparents: i.e. his grandmother had two parents; however, his grandfather had only one parent. 4. How many greatgreatgrand parents did the malebee have? It is possible to make the following table, which sets forth a family tree of each male bee and each female bee. Chapter 2 Fibonacci and Lucas Numbers 67 It follows from Table 2.2 that the reproduction of honeybees is carried out according to “Fibonacci’s principle!” Table 2.2. Fibonacci numbers in a family tree of honeybees Number of malebees Number of female bees Number of parents Number of grand parents Number of greatgrand parents Number of greatgreat grand parents Number of greatgreat greatgrand parents 1 2 3 5 8 2 3 5 8 13 2.3. Numerology and Fibonacci Numbers 2.3.1. Some Mathematical Properties of Numerological Values Increasingly popular today, numerology is an ancient tradition that attract ed attention of civilization’s greatest minds. There is evidence that numerology was in use in China, Greece, Rome, and Egypt for a long time before Pythago ras, who is generally considered to be the “father” of numerology. Numerology, not unlike astrology and other esoteric subjects, has experi enced a revival in recent years. Many modern numerologists have adopted the original Pythagorean system. It is a simple system that assigns a number value (from one to nine) to every letter of the alphabet: A is 1, B is 2, and so on. According to traditional summation, we add for example, 7+8=15. In nu merology, this is not enough: we use a further method of summation known as the “Fadic” system, or “natural summation.” This simply means we sum two or more digits together until we arrive at a single digit. For our example, 7 and 8 add up to 15, however, in numerology we sum 1+5, for a final answer 6. Despite its esoteric aspects, numerology involves a deep mathematical conception. In fact, a calculation of the numerological values of one or anoth er number from the mathematical point of view is very close to obtaining the remainder of some number through division by 9. In mathematics such oper ation is called a reduction by modulo 9. Let us denote a numerological value of the integer a by k(a). We say that two integers a and b are comparable by modulo m (m being a natural number) Alexey Stakhov THE MATHEMATICS OF HARMONY 68 if at the division by m they result in the same remainder. In other words, a and b are comparable by modulo m if their remainder a and b are identical. Example: 32 and 39 are comparable by modulo 7 because 32=7×4+4, 39=7×5+4, both having the same remainder: 4. The statement “a and b are comparable by modulo m” is written in the form: a≡b (mod m). Note that there is only one distinction between the numerological value of the number a and its remainder by modulo m=9. Consider, for example, the Fibonacci number of 144. As this number is perfectly divisible by 9, 144≡0 (mod 9) and the numerological value of the number 144 is equal to 9, i.e. k(144)=1+4+4=9. A relation of comparison by the modulo m has many properties similar to the properties of traditional integers. For example, if a1≡b1 (mod m) and a2≡b2 (mod m), then we have a1a2≡b1b2 (mod m) and (2.5) a1+a2≡(b1+b2)(mod m). As 10≡1 (mod 9) then by using (2.5) it is easy to prove that (2.6) (10i)≡1 (mod 9). Now, let us represent a natural number N in the decimal system: (2.7) (2.8) N = bn 10n −1 + bn −110n −2 + ... + bi 10i −1 + ... + b2 101 + b1100 , where bn,bn1,…,bi,…,b2,b1 are decimal numerals of the number N that take their values from the set of the decimal numerals {0,1,2,3,4,5,6,7,8,9}. The abridged notation of the sum (2.8) N= bn,bn1,…,bi,…,b2,b1 is called a decimal notation of the number N. By using (2.5) through (2.8), we can write: k ( N ) = k ( bn + bn −1 + ... + bi + ... + b2 + b1 ) . If we represent the sum (2.9) S = bn + bn −1 + ... + bi + ... + b2 + b1 in the decimal system (2.8) S = d m 10m −1 + d m −110m −2 + ... + d j 10 j −1 + ... + d2 101 + d1100 , (2.10) then we can obtain a new expression for the numerological value of the number N: ( ) k ( N ) = k dm + dm −1 + ... + d j + ... + d2 + d1 . (2.11) Chapter 2 Fibonacci and Lucas Numbers 69 This procedure continues until the last sum of the kind (2.11) is less than the number 10. It is the numerological value of the number N. From the above examination, it is easy to deduce a number of mathemat ical properties of the “numerological values.” First of all, we can find a numer ological value of the sum of two numbers, for example, N1+N2. For this pur pose, we represent the numbers N1 and N2 in the decimal notation (2.8): (2.12) N 2 = d n 10n −1 + d n −110n − 2 + ... + d i 10i −1 + ... + d2 101 + d1100 Then, we can represent the sum N1+N2 as follows: N1 + N 2 = cn 10n −1 + cn −110n −2 + ... + ci 10i −1 + ... + c2 101 + c1100 (2.13) (2.14) + dn 10n −1 + d n −110n −2 + ... + d i 10i + ... + d2 10i −1 + d1100 . It follows from (2.14) that the numerological value of the sum N1+N2 is calculated according to the expression: k ( N 1 + N 2 ) = k ( cn + cn −1 + ... + ci + ... + c2 + c1 ) + dm + dm −1 + ... + d j + ... + d2 + d1 (2.15) = k ( cn + cn −1 + ... + ci + ... + c2 + c1 ) + k d m + d m −1 + ... + d j + ... + d2 + d1 . ( ( ) ) We can write the expression (2.15) as follows: k ( N1 + N 2 ) = k ( N1 ) + k ( N 2 ) . (2.16) Thus, the numerological value of the sum N1+N2 is equal to the sum of the numerological values of the initial numbers N1 and N2 . As examples, we calcu late the numerological value of the sum: 17711+5702887=5720598. It is clear that the numerological value of the numbers 17711, 5702887 and 5720598 are equal, respectively, k(17711)=1+7+7+1+1=17, 1+7=8; k(5702887)=5+7+0+2+8+8+7=37, 3+7=10, 1+0=1; k(5720598)= 5+7+2+0+5+9+8=36, 3+6=9. On the other hand, we have k(5720598)=k(17711)+k(5702887)=8+1=9. 2.3.2. Fibonacci Numerological Series Let us now examine the Fibonacci numbers from numerological points of view. For this purpose we write the first 48 Fibonacci numbers together with their numerological values k(F1) through k(F48) (Table 2.3). THE MATHEMATICS OF HARMONY Alexey Stakhov 70 Table 2.3. Numerological values of Fibonacci numbers n 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Fn 1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181 6765 10946 17711 28657 46368 k(Fn) n Fn k(Fn) 1 1 2 3 5 8 4 3 7 1 8 9 8 8 7 6 4 1 5 6 2 8 1 9 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 75025 121393 196418 317811 514229 832040 1346269 2178309 3524578 5702887 9227465 14930352 24157817 39088169 63245986 102334155 165580141 267914296 433494437 701408733 1134903170 1836311903 2971215073 4807526976 1 1 2 3 5 8 4 3 7 1 8 9 8 8 7 6 4 1 5 6 2 8 1 9 Let us consider the recursive relation (2.3) for Fibonacci numbers. If we use a general identity (2.16), we can write the following recursive relation for the numerological values of the adjacent Fibonacci numbers: k ( Fn ) = k ( Fn −1 ) + k ( Fn −2 ) . (2.17) We can see from Table 2.3 that this regularity is valid for all Fibonacci num bers in that table. For example, by using Table 2.3, we can write: k(F41) = 4. We can also calculate the numerological value of the Fibonacci number F41 by using the recursive relation (2.17). In fact, k ( F41 ) = k ( F40 ) + k ( F39 ) = 6 + 7, 1 + 3 = 4. Analysis of the numerological values k(F1) to k(F48) lead us to an unexpect ed result. Beginning with the Fibonacci number F25, the numerological values start to recur, that is, the numerological values k(F25) to k(F48) are a repetition of the numerological values k(F1) to k(F48). We can continue Table 2.3 in order to be convinced that the numerological values k(F49) to k(F72) are a repetition of k(F1) to k(F24) and k(F25) to k(F48). Thus, the Fibonacci Numerological Series k(Fi) (i=1, 2, 3, ...) is a periodic sequence with the period of length 24: Chapter 2 Fibonacci and Lucas Numbers 71 1, 1, 2, 3, 5, 8, 4, 3, 7, 1, 8, 9, 8, 8, 7, 6, 4, 1, 5, 6, 2, 8, 1, 9 (2.18) Let us divide the period (2.18) into two parts (with 12 numbers in each part): 1 1, 1, 2, 3, 5, 8, 4, 3, 7, 1, 8, 9 2 8, 8, 7, 6, 4, 1, 5, 6, 2, 8, 1, 9 and then sum them in pairs. As a result, we obtain the regularity presented in Table 2.4. It follows from Table 2.4 that the sum of the first 11 pairs of the Fibonacci numerological series is equal to the number 9. However, the sum of the last, that is, the 12th pair is equal to 9+9 = 18; note that the numerological value of this last sum is also equal to 9. If we sum all numbers of the first period, then this sum is equal to 1+1+2+3+5+8+4+3+7+1+8+9+8+8+7+6+4+1+5+6+2+8+1+9 = 117. Table 2.4. Regularity The numerological value of the sum 117 is again of the Fibonacci equal to 9. This means that the numerological value of numerological series the sum of the first 24 Fibonacci numbers is equal to 9. By analogy, we can assert that the numerological sum 1 + 8 = 9 1 + 8 = 9 of the next 24 Fibonacci numbers will also be equal to 2 + 7 = 9 9. Thus, if we begin from the first Fibonacci numbers F1 3 + 6 = 9 to F24 and examine the subsequent periods of 24 Fi 5 + 4 = 9 bonacci Numbers, for example, F25 to F48, F49 to F72, and 8 + 1 = 9 4 + 5 = 9 so on, then we will find that the numerological values 3 + 6 = 9 of these sums are each equal to 9. In order to calculate 7 + 2 = 9 the numerological value of the Fibonacci series, we 1 + 8 = 9 should calculate the numerological value of the follow 8 + 1 = 9 9 + 9 = 18 ing infinite sum 9+9+…+9+…. It is easy to prove that the numerological value of this sum is equal to 9. This reasoning results in the following theorem. Theorem 2.1. The Fibonacci numerological series has the period (2.18) of length 24; here the numerological value of the sum of the first 24 Fibonacci num bers and all subsequent periods of 24 Fibonacci numbers are each equal to 9. This means that the number 9 is a “numerological essence” of the Fi bonacci series, that is, this number expresses a “sacred” property of the Fibonacci series. 2.3.3. Another Periodic Properties of Fibonacci Numbers It is necessary to point out that the period (2.18) that appears in the nu merological study of Fibonacci numbers is not the only example of a periodic THE MATHEMATICS OF HARMONY Alexey Stakhov 72 property of Fibonacci numbers. For example, we can calculate the values of the Fibonacci series by Mod 2 and Mod 3 (see Table 2.5). Table 2.5. Periodicities of Fibonacci numbers taken by Mod 2 and Mod 3 Fn 1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 Mod 2 1 1 0 1 1 0 1 1 0 1 1 0 1 1 0 1 Mod 3 1 1 2 0 2 2 1 0 1 1 2 0 2 2 1 0 We can see that the Fibonacci numbers by Mod 2 have a period 1, 1, 0 of length 3 and by Mod 3 a period 1, 1, 2, 0, 2, 2, 1, 0 of length 8. By studying the numerical sequences obtained by taking Fibonacci series by Mod k, the Amer Table 2.6. Periodicities of Fibonacci numbers by different k module Modulo k Length of the period 2 3 4 5 6 7 8 9 12 16 3 8 6 20 24 16 12 24 24 24 ican scientist Jay Kappraff found the following periodicities of such sequenc es as given in Table 2.6. Note that mathematicians discovered many interesting numbertheoretic properties of Fibonacci numbers concerning their divisibility. Consider the following remarkable properties of Fibonacci numbers proved in [13]: 1. If n is divisible by m, then Fn is divisible by Fm. For example, consider two Fibonacci numbers, F26=121393 and F13=233. We can see that F26 is divis ible by F13 (121393:233=521) because 26 is divisible by 13 (26:13=2). 2. If Fn≡0 (mod k) and Fm≡0 (modk) then also Fn=m≡0 (mod k) and Fnm≡0 (mod k) (for n>m). For example, consider two Fibonacci numbers, F24=46368 and F12=144. We can calculate that F24≡0 (mod 9) and F12≡0 (mod 9); then for the Fibonacci numbers F24+12=F36=14930352 and F2412=F12=144 we have: F36≡0 (mod 9) and F12≡0 (mod 9). A surprising periodicity of the Fibonacci series taken by Mod k (see Ta bles 2.4 2.6) produces a feeling of rhythm and harmony hidden in this re markable numerical sequence! 2.4. Variations on Fibonacci Theme The variations on a given theme in music are known as genre. A distinctive feature of the musical works of various genres consists of the fact that they be Chapter 2 Fibonacci and Lucas Numbers 73 gin, in most cases, with one simple essential musical theme, which thereinafter undergoes considerable changes of tempo, mood and nature. However, no mat ter how extreme the variations are, the listeners are absolutely impressed that each variation is a natural development of the main theme. If we follow the example of musical genre and select a simple mathemati cal subject, such as the Fibonacci series, we can consider this series together with its numerous variations. 2.4.1. Formulas for the Sums of Fibonacci Numbers Fibonacci numbers have a number of delightful mathematical properties that have stimulated the imaginations of mathematicians over the centuries. Let us calculate, for example, the sum of the first n Fibonacci numbers. Begin from the simplest sums: 1+1 = 2 = 3 −1 1+1+ 2 = 4 = 5 −1 1+1+ 2 + 3 = 7 = 8 −1 1 + 1 + 2 + 3 + 5 = 12 = 13 − 1 (2.19) If we consider in the sums (2.19) the numbers marked by bold type: 3, 5, 8, 13, …, then it is easy to see that they are Fibonacci numbers! Then, we can write the sums (2.19) as follows: F1 + F2 = F4 − 1; F1 + F2 + F3 = F5 − 1; F1 + F2 + F3 + F4 = F6 − 1. It is clear that the general formula has the following form: (2.20) F1 + F2 + ... + Fn = Fn + 2 − 1. Now, let us consider the sum of the n sequential Fibonacci numbers with the odd indexes 1,3,5,…,2n1,…. To this end we start from the simplest sums: 1+ 2 = 3 1+ 2 + 5 = 8 1 + 2 + 5 + 13 = 21 (2.21) 1 + 2 + 5 + 13 + 34 = 55 By analyzing (2.21), we find the following regularity for Fibonacci num bers: the sum of the n sequential Fibonacci numbers with the odd indexes is equal to Fibonacci numbers! In general, the partial sums (2.21) can be written in the form of the following general identity: F1 + F3 + F5 + ... + F2n −1 = F2n . (2.22) Alexey Stakhov THE MATHEMATICS OF HARMONY 74 It is easy to prove the similar formulas for the sums of the n sequential Fibonacci numbers with even indexes: (2.23) F2 + F4 + F6 + ... + F2n = F2n +1 − 1. Now, let us find the sum of the squares of the n sequential Fibonacci numbers: F12 + F22 + ... + Fn2 . Start from the analysis of the simplest sums of the kind (2.24): (2.24) 12 + 12 = 2 = 1 × 2 12 + 12 + 22 = 6 = 2 × 3 12 + 12 + 22 + 32 = 15 = 3 × 5 (2.25) 12 + 12 + 22 + 32 + 52 = 40 = 5 × 8 The analysis of (2.25) results in the following general formula: (2.26) F12 + F22 + ... + Fn2 = Fn Fn +1 , that is, the sum of squares of n sequential Fibonacci numbers is equal to the product of the greatest Fibonacci number used in this sum multiplied by the next Fibonacci number! Also we can find the sum of the squares of two adjacent Fibonacci numbers: Fn2 + Fn2+1 . (2.27) Start from the analysis of the simplest sums of the kind (2.27): 1 + 12 = 1 + 1 2 = 2 12 + 22 = 1 + 4 = 5 22 + 32 = 4 + 9 = 13 (2.28) 32 + 52 = 9 + 25 = 34 By analyzing (2.28), we can find another remarkable regularity: the sum of squares of two adjacent Fibonacci numbers is always equal to a Fibonacci number! The general form of this regularity can be expressed as follows: Fn2 + Fn2+1 = F2n +1 . (2.29) 2.4.2. Connection of Fibonacci Numbers to the Golden Mean Now, let us show a connection of Fibonacci numbers to the golden mean. With this purpose in mind, we can examine a sequence of fractions that are built up by the ratios of adjacent Fibonacci numbers: Chapter 2 Fibonacci and Lucas Numbers 75 1 2 3 5 8 13 21 34 , , , , , , , , ... (2.30) 1 1 2 3 5 8 13 21 The first terms of the sequence (2.30) have the following values: 1 2 3 5 8 13 21 = 1; = 2; = 1.5; = 1.666; = 1.6; = 1.625; = 1.615; ... 1 1 2 3 5 8 13 The question is: what is the limit of the sequence (2.30) if we direct the index n to infinity? To answer this question, let us consider the representation of the golden mean in the form (1.15). It is easy to prove that the sequence (2.30) is connected directly with the representation (1.15). In fact, the fractions (2.30) are sequential approximations of the continuous fraction (1.15), namely: 1 =1 1 (the first approximation); 2 1 =1+ 1 1 (the second approximation); 3 1 =1+ 1 2 1+ 1 (the third approximation); 5 =1+ 3 1 1+ (the fourth approximation). 1 1+ 1 1 If we continue this process to infinity, we obtain: lim Fn n →∞ F n −1 =τ= 1+ 5 . 2 (2.31) The result given by the expression (2.31) is the “key” result for our re search because it shows a deep connection between the Fibonacci numbers and the golden mean. This means, that like the golden mean itself, Fibonacci numbers express harmony in the world around us! 2.4.3. “Iron Table” by Steinhaus Renowned Polish mathematician Steinhaus constructed a table of ran dom numbers by using the golden mean. For this purpose, he multiplied 10 000 integers from 1 to 10 000 by the number ϕ=τ1=0.61803398, where τ is Alexey Stakhov THE MATHEMATICS OF HARMONY 76 the golden mean. As a result, he obtained the sequence of numbers multiplied by ϕ, that is: 1ϕ, 2ϕ 3ϕ, ... , 4181ϕ, ... , 6765ϕ, ... , 10000. Steinhaus called this numerical sequence the Golden Numbers. Each Golden Number consists of integer and fractional parts. For example, the number 1000ϕ = 618.03398 has the integer part 618 and the fractional part 0.03398. The number 4181ϕ=2584.00007 has the integer part 2584 and the fractional part 0.0007, and so on. Moreover, neither a Golden Number with fractional part equal to 0 exists nor do two Golden Numbers with the equal fractional parts. This means that every Golden Number has a unique fractional part. If we put the Golden Numbers in order in a special table according to their increasing fractional parts, we find that the number 4181ϕ has the least frac tional part and, therefore, this Golden Number should start this table; also we find that the number 6765ϕ has the largest fractional part and, therefore, this Golden Number has to end the table: 4181 3194 …… 8739 0987 8362 7365 …… 1974 5168 1597 0610 …… 6155 9349 5778 4791 …… 3571 2584 9959 8972 …… 7752 6765 Steinhaus named this the Iron Table taking into consideration some of its unique properties. The Iron Table demonstrates deep connections with Fibonac ci numbers. The first property is that a difference between the adjacent Gold en Numbers of the Iron Table by absolute value is always equal to one of three numbers: 4181, 6765 and 2584. In fact, we have: 8362 − 4181 = 4181, 8362 − 1597 = 6765, 5778 − 1597 = 4181, ... 9349 − 5168 = 4181, 9349 − 2584 = 6765, 6765 − 2584 = 4181. It is easy to find the numbers of 2584, 4181 and 6765 if we continue the Fibonacci sequence: 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 897, 1597, 2584, 4181, 6765,... Hence, the numbers 2584, 4181, 6765 are three adjacent Fibonacci numbers: F18 = 2584, F19 = 4181, F20 = 6765. We can see that the Iron Table starts with the Fibonacci number F18=2584 and ends with the Fibonacci numbers F19=4181 and F20=6765. Chapter 2 Fibonacci and Lucas Numbers 77 It is clear that the Iron Table may be constructed for any arbitrary natural number N. Polish scientist Jan Grzedzielski in his book EnergyGeometric Code of Nature [26] analyzed the Iron Tables for the cases N=Fn where Fn is Fibonac ci number. He obtained an interesting regularity that appears at the conver sion of the Iron Table with N=Fn1 into the next Iron Table with N=Fn. This latter table is constructed from the preceding table (with N=Fn1) by means of the disposition of new numbers Fn1+1, Fn2+2,…,Fn1,Fn on particular positions in the new Iron Table. In Grzedzielski’s opinion, this method of Iron Table con struction “resembles the functioning of all radiation spectra in Nature.” 2.5. Lucas Numbers 2.5.1. FrancoisEdouardAnatole Lucas Fibonacci did not continue to study the mathematical properties of the nu merical series (2.4). However, the study of Fibonacci numbers was continued by other mathematicians. Since the 19th century, the mathematical works de voted to Fibonacci numbers, according to the witty expression of one mathe matician, “began to be reproduced like Fibonacci’s rabbits.” The French math ematician Lucas became one of the leaders of this research in the 19th century. What do we know about Lucas? The French math ematician FrancoisEdouardAnatole Lucas was born in 1842. He died in 1891 as a result of an accident that occurred at a banquet, when a dish was smashed and a splinter wounded his cheek. Lucas died from an infec tion some days later. Lucas’ major works fall in the area of number theory and indeterminate analysis. In 1878 Lucas gave the criterion for the determination of the “pri mality” of Mersenne numbers of the kind Mp=2p1. FrancoisEdouard Anatole Lucas Applying his own method, Lucas proved that the (18421891) Mersenne number M127 = 2127 − 1 = 170141183460469231731687303715884105727 is prime. For 75 years, this number was the greatest prime number known in mathematics. He also found the 12th “perfect number” and formulated a num ber of interesting mathematical problems. Alexey Stakhov THE MATHEMATICS OF HARMONY 78 We can give some explanation for Lucas’ scientific achievements. It is well known that the prime numbers are divisible only by 1 and itself. The first few prime numbers are: 2, 3, 5, 7, 11, 13, … . Already the Pythagoreans knew that the number of prime numbers is infinite (the proof of this fact is in The Ele ments of Euclid). A study of the prime numbers and determination of their distribution in the natural series is a rather difficult problem in number theo ry. Therefore, the scientific result that was obtained by Lucas in the field of prime numbers, undoubtedly, belonged to the category of outstanding mathe matical achievements. It is curious, that Lucas took a great interest in the socalled Perfect Num bers. What are Perfect Numbers? As is well known, the Pythagorean theory of numbers had a qualitative character, that is, the Pythagoreans were interest ed in the qualitative aspects of numbers. They attributed to numbers some unusual properties. In this connection, the socalled Perfect Numbers are of special interest. For example, consider the number 6. Its feature is that this number is equal to the sum of its divisors, that is, 6=1+2+3. Besides the num ber 6, the Pythagoreans knew another two Perfect Numbers, 28 and 496: 28=1+2+4+7+14; 496=1+2+4+8+16+31+62+124+248. The fourth perfect number is 8128. It was also known by ancient math ematicians. It is proven, that in the process of movement along natural series the Perfect Numbers are found less frequently. Only four Perfect Numbers (4, 28, 496 and 8128) are found in the first 10,000 numbers of the natural series. The search for Perfect Numbers became a fascinating passion for many math ematicians. The fifth Perfect Number 212(2131) was found in the 15th centu ry by the German mathematician Regiomontan. In the 16th century, the German scientist Sheybel found two new Perfect Numbers: 8589869056 and 137438691328. Lucas in the 19th century found the twelfth Perfect Number. Research in this area continues today where all the power of modern com puters is being used. For example, the 18th Perfect Number that was found by means of computer modeling has 2000 decimal digits. In honor of Lucas, it is necessary to note one more of his scientific predic tions. Already in the 19th century, long before the occurrence of modern com puters, Lucas paid attention to technical advantages of the binary notation for technical realization of computers and machines. This means that he an ticipated by a century the outstanding American physicist and mathemati cian John von Neumann’s preference for the binary system in the technical realization of electronic computers (John von Neumann’s principles). Chapter 2 Fibonacci and Lucas Numbers 79 2.5.2. Some Properties of Lucas Numbers However, it is most relevant for our book that in the 19th century Lucas attracted the attention of mathematicians to the remarkable numeric sequence 1, 1, 2, 3, 5, 8, 13, 21, 34, …, which he named Fibonacci Numbers. In addition, Lucas introduced the concept of generalized Fibonacci numbers that are cal culated according to the following general recursive relation: Gn = Gn −1 + Gn −2 (2.32) at the initial terms (seeds) G1 and G2. For example, the sequence of numbers 2, 8, 10, 18, 28, 46, … falls into the generalized class of Fibonacci numbers satisfying the recursive relation (2.5) at the seeds G1=2 and G2=8. However, the main numerical sequence of the type (2.32), introduced by Lucas in the 19th century, is a numerical sequence given by the following re cursive relation: Ln = Ln −1 + Ln −2 (2.33) at the seeds L1 = 1, L2 = 3. (2.34) Then, by using the recursive relation (2.33) at the seeds (2.34) we can calculate the following numerical sequence called Lucas Numbers: 1, 3, 4, 7, 11, 18, 29, 47, 76, 123, 199, …. (2.35) If we make the reasonings for Lucas numbers similar to Fibonacci num bers, we can prove the following identities: L1 + L2 + ... + Ln = Ln + 2 − 3 L1 + L3 + L5 + ... + L2n −1 = L2n − 2 L1 + L3 + L5 + ... + L2n −1 = L2n − 2 L21 + L22 + L23 + ... + L2n = Ln Ln +1 − 2 L2n + L2n +1 = 5F2n +1 lim Ln n →∞ L n −1 =τ= 1+ 5 2 (2.36) (2.37) 2.5.3. The “Extended” Fibonacci and Lucas Numbers Up until now, we have studied Fibonacci and Lucas series Fn and Ln with the positive indices n, that is, n=1,2,3,…. However, they can be extended with Alexey Stakhov THE MATHEMATICS OF HARMONY 80 the negative values of the indices n, that is, when the indices n take their val ues from the set: n=0,1,2,3,…. The “extended” Fibonacci and Lucas numbers are represented in Table 2.7. It follows from Table 2.7. “Extended” Fibonacci and Lucas numbers Table 2.7 that the 0 1 2 3 4 5 6 7 8 9 10 n elements of the “ex 0 1 1 2 3 5 8 13 21 34 55 Fn tended” numerical 0 1 1 2 3 5 8 13 21 34 55 Fn sequences F n and 1 3 4 7 11 18 29 47 76 123 Ln have a number of Ln 2 remarkable mathe 1 3 4 7 11 18 29 47 76 123 Ln 2 matical properties. For example, for the odd indices n=2k+1 the elements of the sequences Fn and Fn coincide, that is, F2k+1= F2k1, and for the even indices n=2k they are oppo site by a sign, that is, F2k=F2k. For the Lucas numbers Ln all is vice versa, that is, L2k=L2k; L2k+1= L2k1. Now, let us consider the numerical sequences given in Table 2.7. Take, for example, the Lucas number L4=7 and compare it with Fibonacci numbers. It is easy to find that L4=7=5+2. However, the numbers 2 and 5 are the Fibonac ci numbers F3=2 and F5=5. Is this a mere coincidence? By continuing examination of Table 2.7, we can find the following correlations that connect Fibonacci and Lucas num bers: 1=0+1, 3=1+2, 4=1+3, 7=2+5, 11=3+8, 18=5+13, 29=8+21, and so on. Now, let us compare the numerical sequences Ln and Fn. Here we find the same regularity, that is, 1=0+(1), 3=1+2, 4=(1)+(3), and so on. Thus, we found the following surprising mathematical identity that connects Lucas and Fibonacci numbers: (2.38) Ln = Fn + Fn +1 , where the index n takes the following values: n= 0, ±1, ±2, ±3, …. If we continue examination of Table 2.7, we can find other interesting iden tities for Fibonacci and Lucas numbers: Ln = Fn + 2Fn −1 (2.39) (2.40) Ln + Fn = 2Fn −1 . Note that the above formulas for Fibonacci and Lucas numbers can be considered to be the “golden treasure” of mathematics! And, by comprehend ing these formulas, we can understand the delight of many outstanding math ematicians of the 20th century, in particular, the Russian mathematician Ni kolay Vorobyov and the American mathematician Verner Hoggatt. It was Pro fessor Nikolay Vorobyov who was the author of the remarkable brochure Chapter 2 Fibonacci and Lucas Numbers 81 Fibonacci Numbers [13] that became a mathematical bestseller of the 20th cen tury. Professor Verner Hoggat was the founder of the Fibonacci Association, the mathematical journal The Fibonacci Quarterly and the author of the book Fibonacci and Lucas Numbers [16]. They saw in Fibonacci and Lucas numbers a “mathematical secret of nature.” The desire to uncover this “secret” inspired them to study these unique mathematical sequences! 2.5.4. Other Remarkable Identities for Fibonacci and Lucas Numbers The next group of formulas is based on the following identity that connects the generalized Fibonacci numbers Gn with the classical Fibonacci numbers Fn: Gn + m = Fm −1Gn + FmGn +1 . (2.41) We can prove this formula by induction on m. For the cases m=1 and m=2 this formula is valid because Gn +1 = F0Gn + F1Gn +1 = Gn +1 and Gn + 2 = F1Gn + F2Gn +1 = Gn + Gn +1 . The basis of the induction is proved. Let us suppose that the formula (2.41) is valid for the cases m=k and m=k+1, that is, Gn + k = Fk −1Gn + FkGn +1 , Gn + k +1 = FkGn + Fk +1Gn +1 . By summarizing these formulas termwise, we obtain the following identity: Gn + k + 2 = Fk +1Gn + Fk + 2Gn +1 . The identity (2.41) is proved. A number of interesting identities for Fibonacci and Lucas numbers fol low from the identity (2.41). Suppose that Gi=Fi and m=n+1. Then, the iden tity (2.41) is reduced to the following: F2n +1 = Fn2+1 + Fn2 . (2.42) For the case Gi=Fi and m=n, the formula (2.41) is reduced to the following: F2n = Fn −1 Fn + Fn Fn +1 = Fn ( Fn −1 + Fn +1 ) = Fn Ln . (2.43) By using the formula (2.43), we can write: Fn +1 Ln +1 − Fn Ln = F2n + 2 − F2n = Fn Ln . (2.44) 2.5.5. Lucas Numbers and Numerology As was mentioned above, the Fibonacci series generates a periodic Fibonac ci numerological series with the period (2.18). There is a question: whether there is a similar periodicity for the Lucas series? Table 2.8 presents 48 Lucas numbers and their numerological values. THE MATHEMATICS OF HARMONY Alexey Stakhov 82 Table 2.8. Numerological values of Lucas numbers n Ln 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 3 4 7 11 18 29 47 76 123 199 322 521 843 1364 2207 3571 5778 9349 15127 24476 39603 64079 103682 k(Ln) n Ln k(Ln) 1 3 4 7 2 9 2 2 4 6 1 7 8 6 5 2 7 9 7 7 5 3 8 2 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 167761 271443 439204 710647 1149851 1860498 3010349 4870847 7881196 12752043 20633239 33385282 54018521 87403803 141422324 228826127 370248451 599074578 969323029 1568397607 2537720636 4106118243 6643838879 10749957122 1 3 4 7 2 9 2 2 4 6 1 7 8 6 5 2 7 9 7 7 5 3 8 2 The analysis of Table 2.8 shows that the numerological Lucas series {k(Ln, n=1, 2, 3,…)} possess mathematical properties similar to the numerolog ical Fibonacci series, that is, the numerological Lucas series is periodic with the period of length 24 of the following kind: 1, 3, 4, 7, 2, 9, 2, 2, 4, 6, 1, 7, 8, 6, 5, 2, 7, 9, 7, 7, 5, 3, 8, 2 (2.45) If we divide the period (2.45) into two parts (by 12 numbers in each part) 1 1, 3, 4, 7, 2, 9, 2, 2, 4, 6, 1, 7 2 8, 6, 5, 2, 7, 9, 7, 7, 5, 3, 8, 2 and then sum them in pairs, we find the following regularity (see Table 2.9): That is, the sums of all 11 mutual numerological values of the period (2.45) are equal to 9 and only one to the sum 9+9=18; however, its numerological value is equal to 9. If we use the same reasoning as for Fibonacci numbers, we can prove the following theorem. Theorem 2.2. The Lucas numerological series has the period (2.45) of length 24; here the numerological value of the sum of the first 24 Lucas numbers and all the next sequential 24 Lucas numbers are equal to 9. Chapter 2 Fibonacci and Lucas Numbers 83 This means that the number 9 is the “numerological essence” of the Lucas series, that is, it expresses some “sa cred” property of Lucas numbers. It is easy to prove that the generalized Fibonacci numbers given by (2.32) has similar general property (see Theorem 2.3). Theorem 2.3. The numerological series of the gen eralized Fibonacci numbers given by the recursive rela tion (2.32) has the period of length 24; here the numer ological value of the sum of the first 24 generalized Fi bonacci numbers and all the next sequential 24 general ized Fibonacci numbers are equal to 9. Table 2.9. Regularity of the Lucas numero logical series 1 + 8 = 9 3 + 6 = 9 4 + 5 = 9 7 + 2 = 9 2 + 7 = 9 9 + 9 = 18 2 + 7 = 9 2 + 7 = 9 4 + 5 = 9 6 + 3 = 9 1 + 8 = 9 7 + 2 = 9 2.6. Cassini Formula 2.6.1. Great Astronomer Giovanni Domenico Cassini Cassini is the name of a famous dynasty of French as tronomers. Giovanni Domenico Cassini (16251712) is the most famous of them and the founder of this dynasty. The following facts illustrate his outstanding contribution to as tronomy. The name Cassini was given to numerous astro nomical objects, the “Cassini Crater” on the Moon, the “Cassini Crater” on Mars, “Cassini Slot” in Saturn’s ring, and “Cassini Laws” of the Moon’s movement. However, the name of Cassini is widely known not only Giovanni Domenico in astronomy, but also in mathematics. Cassini developed a Cassini (16251712) theory of the remarkable geometrical figures known under the name of Cassini Ovals. The mathematical identity that connects three adjacent Fibonacci numbers is well known under the name Cassini Formula. Below we will review this famous formula. Giovanni Cassini was born on June, 8, 1625 in the Italian town of Perinal do. He got his education in Jesuit collegiums in Genoa. During 16441650 he worked in the observatory located near Bologna. In 1650 Cassini took a professorial chair in mathematics and astronomy at the University of Bologna. Cassini’s basic scientific works concerned obser vational astronomy and he became famous, first of all, as a talented observer. Working in Bologna, for the first time in history, he executed numerous posi Alexey Stakhov THE MATHEMATICS OF HARMONY 84 tional observations of the Sun with meridian tools. On the basis of these ob servations he made new Solar tables published in 1662. Owing to Cassini’s researches, the Parisian Meridian was established. Cassini’s work provided the possibility for the creation of the wellknown map of France, the Cassini Map. The glory of Cassini as an astronomer was so great that in 1669 he was elected a member of the Parisian Academy of Sciences. According to Pikar’s recommendation, King Louis XIV invited Cassini to take a position as Direc tor of Parisian observatory. France became his second native land up to the end of his life. In Paris, Cassini made a number of outstanding astronomical discoveries. During 16711684 he discovered several satellites of Saturn. In 1675 he found that the ring of Saturn consists of two parts divided by dark strip called the Cassini Slot. During 16711679 he observed details of the Lu nar surface and in 1679 he made a finer map of the Moon. In 1693 Cassini formulated three empirical laws for the Moon’s movement called Cassini Laws. Cassini died in Paris on September 14, 1712 at the age of 87, absolutely blind but a highly honored man. 2.6.2. Cassini Formula for Fibonacci Numbers The history of science is silent as to why Cassini took such a great interest in Fibonacci numbers. Most likely it was simply a hobby of the Great astron omer. At that time many serious scientists took a great interest in Fibonacci numbers and the golden mean. These mathematical objects were also a hobby of Cassini’s contemporary, Kepler. Now, let us consider the Fibonacci series: 1, 1, 2, 3, 5, 8, 13, 21, 34, … . Take the Fibonacci number 5 and square it, that is, 52=25. Multiply two Fibonacci numbers 3 and 8 that encircle the Fibonacci number 5, so we get 3×8=24. Then we can write: 523×8=1. Note that the difference is equal to (+1). Now, we follow the same process with the next Fibonacci number 8, that is, at first, we square it 82=64 and then multiply two Fibonacci numbers 5 and 13 that encircle the Fibonacci number 8, 5×13=65. After a comparison of the result 5×13=65 with the square 82=64 we can write: 825×13=1. Note that the difference is equal to (1). Further we have: 2 13 8×21=1, 21213×34=1, and so on. Chapter 2 Fibonacci and Lucas Numbers 85 We see that the square of any Fibonacci number Fn always differs from the product of two adjacent Fibonacci numbers Fn1 and Fn+1, which encircle it, by 1. However, the sign of 1 depends on the index n of the Fibonacci number Fn. If the index n is even, then the number 1 is taken with minus, and if odd, with plus. The indicated property of Fibonacci numbers can be expressed by the following mathematical formula: (2.46) Fn2 − Fn −1 Fn +1 = (−1)n +1 . This wonderful formula evokes a reverent thrill, if one recognizes that this formula is valid for any value of n (we remember that n can be some integer in limits from ∞ up to +∞). The alternation of +1 and 1 in the expression (2.45) at the successive passing of all Fibonacci numbers produces genuine aesthetic enjoyment and a feeling of rhythm and harmony. 2.7. Pythagorean Triangles and Fibonacci Numbers 2.7.1. Pythagorean Theorem The Pythagorean Theorem is probably the bestknown theorem in all of geometry. It is remembered by anyone who studied geometry in second ary school, though having “absolutely forgotten” all mathematics. The es sence of this theorem is extremely simple. The Pythagorean Theorem as serts that the sides a, b and c of a right triangle are connected by the fol lowing formula: (2.47) a2+b2=c2. Despite its ultimate simplicity, the Pythagorean Theorem, in the opinion of many mathematicians, refers to a category of the most outstanding mathe matical theorems. Earlier we mentioned Kepler’s widely known statement con cerning the Pythagorean Theorem and the golden mean. Among all the vast sea of geometrical results and theorems, Kepler chose only two results that he named “treasures of geometry”: the Pythagorean Theorem and the “division in extreme and mean ratio” (the golden mean). 2.7.2. Pythagorean Triangles Amongst the infinite number of rightangled triangles, it is the Pythagorean Triangles, for which the numbers a, b, c in (2.47) are integers, which have al ways caused a special interest. Alexey Stakhov THE MATHEMATICS OF HARMONY 86 Pythagorean triangles also can be referred to as a category of the “treasures of geometry,” and a study of such triangles is one of the most fasci nating pages in the history of mathematics. 5 The right triangle with sides 3, 4 and 5 is the most widely known Pythagorean trian 3 gle (Fig. 2.3). It is called the Sacred or Egyp tian Triangle because it was widely used in Ancient Egyptian culture. We mentioned above 4 that this triangle is the main geometrical idea of the Chephren Pyramid in Giza. For the Egyptian Triangle in Fig. 2.3 the Py Figure 2.3. Sacred or thagorean Theorem (2.47) takes the following form: Egyptian Triangle (2.48) 42+32=52. There is a legend that the identity (2.48) was used by the Egyptian land sur veyors and builders for the definition of a right angle on the earth’s surface. For this purpose, they used a rope of the length 12m, for example (12=3+4+5). The rope was divided by means of special loops into three parts of the lengths 3, 4, 5m, respectively. For the definition of a right angle, the Egyptian land surveyor pulled one of the parts of the rope, for example, of the length 3m, and then fixed it on the ground by using special pegs, inserted in two loops. Then, the rope was pulled by means of the third loop and this loop was fixed by using the third peg. Clearly, the angle that is formed between the two smaller sides of the triangle accurately equals 90°. Tradition has it that at the foundation of the Pyramid a ritual procedure to define the right angles at the base of the Pyramid was carried out by the Pharaoh. 2.7.3. FibonacciPythagorean Triangles There is a question: are there other Pythagorean triangles besides the Egyp tian Triangle? Rasko Jovanovich [139] gives an original answer to this question. First, we try to express the identity (2.47) by Fibonacci numbers begin ning with the first four Fibonacci numbers: (2.49) F1 = 1, F2 = 1, F3 = 2, F4 = 3. Perform the following procedure: 1. Multiply the two middle Fibonacci numbers of the series (2.49): F2×F3=1×2=2. 2. Double this result: 2×(F2×F3)=2×2=4. The product 4 is the length of the side a of the Egyptian Triangle in Fig. 2.3, that is, a=2×(F2×F3)=4. 3. Multiply the two external numbers of the series (2.49): F1×F4=1×3=3. Chapter 2 Fibonacci and Lucas Numbers 87 The product 3 is the length of the side b of the Egyptian Triangle in Fig. 2.3, that is, b= F1×F4=3. 4. The third, the longest side c of the Egyptian Triangle, is determined if we add the squares of the interior numbers of the series (2.49), that is, the numbers (F2)2=12=1 and (F3)2=22=4; then their sum is equal to 1+4=5. Hence, c=(F2)2+(F3)2=5. Therefore, we can represent a fundamental identity for the Egyptian Tri angle (2.48) by Fibonacci numbers as follows: 2 2 2 2 2 2 ( F2 × F3 ) + ( F1 × F4 ) = ( F2 ) + ( F3 ) . (2.50) Is this result (2.50) a mere coincidence? To answer this question, we ex amine the next four Fibonacci numbers: (2.51) F2 = 1, F3 = 2, F4 = 3, F5 = 5. If we repeat the above procedure for the series (2.51), we obtain: 1. 2×3=6 2. 2×6=12; hence, a=12 3. 1×5=5; hence, b=5 4. 22=4, 32=9, 4+9=13; hence, c=13 It is easy to be convinced, that the sides a=12, b=5 and c=13 make up a Pythagorean triangle because: 122+52=132. These examples allow us to formulate a general rule to find Pythagorean triangles by using four adjacent Fibonacci numbers: Fn , Fn +1 , Fn + 2 , Fn + 3 . (2.52) Perform the following: 1. Multiply two middle Fibonacci numbers of the series (2.52): Fn +1 × Fn + 2 . 2. Double this result: 2 × ( Fn +1 × Fn +2 ) ; hence, a = 2 × ( Fn +1 × Fn +2 ) . 3. Multiply the two external Fibonacci numbers of the series (2.52): Fn × Fn +3 ; hence, b = Fn × Fn +3 . 4. Square the two interior Fibonacci numbers of the series (2.52), 2 2 2 2 ( Fn+1 ) and ( Fn+2 ) , and then add them: c + ( Fn +1 ) + ( Fn +2 ) . For the general case of (2.52) the main identity for Pythagorean triangles is as follows: ( 2 × Fn +1 × Fn +2 )2 + ( Fn × Fn +3 )2 = ( Fn2+1 + Fn2+2 ) . 2 (2.53) Now let us consider the Pythagorean triangle for the case n=3. For this case, the four adjacent Fibonacci numbers are the following: 2, 3, 5, 8. (2.54) THE MATHEMATICS OF HARMONY Alexey Stakhov 88 Then, according to the above algorithm, the sides of the Pythagorean tri angle can be found in the following manner: a = 2 × 3 × 5 = 30; b = 2 × 8 = 16; c = 32 + 52 = 9 + 25 = 34. The Pythagorean Theorem for this case is: 302+162=342. Finally, for the case n=4 the four adjacent Fibonacci numbers are the follow ing: 3, 5, 8, 13, and the sides of Pythagorean triangle are equal to, respectively, a = 2 × 5 × 8 = 80; b = 3 × 13 = 39; c = 52 + 82 = 35 + 64 = 89. The Pythagorean Theorem for this case is as follows: 802+392=892. Table 2.10 represents Pythagorean triangles for the initial values n. It is essential to Table 2.10. FibonacciPythagorean triangles note that side c of the Pythagorean triangles n Fn Fn+1 Fn+2 Fn+3 a b c of Table 2.10 are calcu 1 1 1 2 3 4 3 5 lated by the formula: 2 1 2 3 5 12 5 13 3 2 3 5 8 30 16 34 4 3 5 8 13 80 39 89 6 8 13 21 34 546 272 610 7 13 21 34 55 1428 715 1597 c = Fn2+1 + Fn2+ 2 . (2.55) By using the identi ty (2.29), we can write: c = F2( n +1) +1 , that is, the hypotenuse c of the FibonacciPy 9 34 55 89 144 9790 4869 10946 thagorean triangle is always equal to some Fibonacci number that is confirmed by Table 2.10. 8 21 34 55 89 3740 1869 4181 2.7.4. LucasPythagorean Triangles The above procedure for the FibonacciPythagorean triangles also proves to be valid for Lucas numbers. For example, the first four adjacent Lucas num bers 1, 3, 4, 7 result in the LucasPythagorean triangle with the sides: a = 2 × 3 × 4 = 24; b = 1 × 7 = 7; c = 32 + 42 = 9 + 16 = 25. For this case the Pythagorean Theorem is as follows: 2 24 +72=252. The next four adjacent Lucas numbers 3, 4, 7, 11 result in the next Lucas Pythagorean triangle with the sides: a = 2 × 4 × 7 = 56; b = 3 × 11 = 33; c = 42 + 72 = 16 + 49 = 65. For this case the Pythagorean Theorem is as follows: 2 56 +332=652. Table 2.11 represents the LucasPythagorean triangles for the initial values of n. Chapter 2 Fibonacci and Lucas Numbers 89 Table 2.11. LucasPythagorean triangles n Ln Ln+1 Ln+2 Ln+3 a b c 1 1 3 4 7 24 7 25 2 3 4 7 11 56 33 65 3 4 7 11 18 154 72 170 4 7 11 18 29 396 203 445 5 11 18 29 47 1044 517 1165 6 18 29 47 76 2726 1368 3050 7 29 47 76 123 7144 3567 7985 8 9 47 76 76 123 123 199 199 322 18696 48954 9353 24472 20905 54730 2.7.5. Fibonacci and Lucas Right Triangles Now, let us consider the identity Fn2 + Fn2+1 = F2n +1 (2.29) as an expression of the Pythagorean Theorem (2.47) for the right triangle with the sides a = Fn ; b = Fn +1 and c = F2n +1 . We name Table 2.12. Fibonacci right triangles these right triangles Fibonacci Right Trian 2 F2 n 1 n F Fn 1 gles because they are based on the impor 1 1 1 2 tant identity (2.29) that connects the adja 2 1 4 5 cent Fibonacci numbers. Note that Fibonacci 3 4 9 13 Right Triangles are not Pythagorean. Nev 4 9 25 34 ertheless, they are a very interesting class 5 25 64 89 6 64 169 175 of right triangles. Fibonacci right triangles 7 169 441 610 are given by Table 2.12. 8 441 1156 1597 Note that similar triangles can also be 9 1156 3025 4181 constructed on the basis of Lucas numbers if we use the identity (2.36) that gives Lucas right triangle with the sides: 2 n Table 2.13. Lucas right triangles n Ln 2 Ln 1 2 5F2 n 1 1 2 3 4 5 6 7 8 9 1 9 16 49 121 324 841 2209 5776 9 16 49 121 324 841 2209 5776 15129 10 25 65 170 445 1165 3050 7985 7305 a = Ln , b = Ln +1 and c = 5F2n +1 . Table 2.13 gives the Lucas right tri angles. The above connection of Fibonacci and Lucas numbers with Pythagorean triangles allows us to propose the existence of an in finite set of Fibonacci and Lucas Pythagore an triangles. This fact is additional testimo ny for the fundamental character of Fi bonacci and Lucas numbers! Alexey Stakhov THE MATHEMATICS OF HARMONY 90 2.8. Binet Formulas 2.8.1. Jacques Philippe Marie Binet In the 19th century, the interest in Fibonacci numbers and the golden mean in science and mathematics again arose. The mathematical works of the French mathematicians Lucas and Binet were a bright reflection of this inter est. We have mentioned above about the French 19th century mathematician Lucas (18421891), who revived the interest in Fibonacci numbers of the sci entific 19th century community. The French mathematician Jacques Philippe Marie Binet was the other 19th century enthusiast of Fi bonacci numbers and the golden mean. He was born on February 2, 1776 in Renje and died on May 12, 1856 in Paris. Following his graduation from the Polytech nic School in Paris in 1806 Binet worked at the Bridge and Road Department of the French government. He became a teacher at the Polytechnic school in 1807 and then Assistant Professor of applied analysis and descriptive geometry. Binet studied the fundamentals Jacques Philippe Marie Binet (17761856) of matrix theory and his work in this direction was continued later by other researchers. He discovered in 1812 the rule for ma trix multiplication, which glorified his name more than any of his other works. Binet worked in other areas in addition to mathematics. He published many articles on mechanics, mathematics and astronomy. In mathematics, Binet introduced the notion of the “beta function”; also he considered the linear difference equations with alternating coefficients and established some metric properties of conjugate diameters, etc. Among his many honors, Binet was elected to the Parisian Academy of sciences in 1843. Binet entered Fibonacci number theory as author of the famous mathe matical formulas called Binet Formulas. These formulas link the Fibonacci and Lucas numbers with the golden mean and, undoubtedly, belong amongst his tory’s most famous mathematical formulas. 2.8.2. Deducing Binet Formulas In order to deduce Binet formulas, we must first consider the remarkable identity that connects the adjacent degrees of the golden mean: Chapter 2 Fibonacci and Lucas Numbers 91 (2.56) τn = τn −1 + τn −2 , where τ is the golden mean and n takes its values from the set: 0, ±1, ±2, ±3, …. Now, let us write the expressions for the zero, first and minusfirst terms of the golden mean: 2 + 0× 5 1 1+ 5 −1 + 5 . , τ = , and τ −1 = (2.57) 2 2 2 Let us remember that the degrees of τ are connected one to the other by the following identity: (2.58) τ2 = τ + 1, that is a partial case of the general identity (2.56) for the case n=2. By employing the expressions of (2.57) and the identities (2.56) and (2.58), we can represent the second, third and fourth degrees of the golden mean as follows: 3+ 5 3 4+2 5 4 7+3 5 ; τ = τ 2 + τ1 = ; τ = τ3 + τ2 = . τ 2 = τ1 + τ0 = (2.59) 2 2 2 Is it possible to see some regularity in the formulas (2.57) and (2.59)? First of all, we can see that each expression for any degree of the golden mean has the following typical form: τ0 = 1 = A+ B 5 . 2 What are the numerical sequences A and B in these formulas? It is easy to see that the series of the numbers А is the number sequence 2, 1, 3, 4, 7, 11, 18, …, and the series of the numbers В is the number sequence 0, 1, 1, 2, 3, 3, 5, 8, …. Thus, the first sequence is Lucas numbers Ln and the second one is Fibonacci numbers Fn! It follows from this reasoning that the general formula that allows representation of the nth degree of the golden mean by Fibonacci and Lucas numbers has the following form: L + Fn 5 τn = n . (2.60) 2 Note that the formula (2.60) is valid for each integer n taking its values from the set 0, ±1, ±2, ±3, …. By using the formula (2.60), it is possible to represent the “extended” Fi bonacci and Lucas numbers by the golden mean. For this purpose, it is enough to write the formulas for the sum or difference of the nth degrees of the gold n −n en mean τn + τ − n and τ − τ as follows: (L + L− n ) + (Fn + F− n ) 5 τn + τ − n = n (2.61) 2 Alexey Stakhov THE MATHEMATICS OF HARMONY 92 ( Ln − L− n ) + ( Fn − F− n ) 5 . (2.62) 2 Now, let us consider the expressions of the formulas (2.61) and (2.62) for the even values of the index n=2k. For this purpose, recall once again the fol lowing wonderful property of Fibonacci numbers: for the even values of n the Fibonacci numbers F2k and F2k are equal by absolute value and opposite by sign, that is, F2k=F2k and likewise for the Lucas numbers L2k and L2k, that is, L2k=L2k. Then for the case of n=2k the (2.61) and (2.62) take the following form: τn − τ − n = τ2 k + τ −2 k = L2 k (2.63) (2.64) τ2 k − τ −2 k = F2 k 5 . For the odd n=2k+1 we have the following relations for the “extended” Fibonacci and Lucas numbers: F2k1=F2k+1 and L2k1=L2k+1. Then, for this case the formulas (2.63) and (2.64) take the following form: τ2 k +1 + τ ( − 2 k +1) = F2 k 5 (2.65) − 2 k +1 (2.66) τ2 k +1 − τ ( ) = L2 k +1 . We can now represent the formulas (2.65) and (2.66) in the following com pact forms: τn + τ− n for n = 2k Ln = n −n τ − τ for n = 2k + 1 (2.67) τn + τ− n for n = 2k + 1 5 . Fn = n −n τ − τ for n = 2k 5 (2.68) The analysis of the formulas (2.67) and (2.68) provide us with “aesthetic pleasure” and once again we are convinced of the power of the human mind! We know that the Fibonacci and Lucas numbers are always integers. On the other hand, any degree of the golden mean is an irrational number. It follows from this that the integer numbers Ln and Fn can be represented with the help of the for mulas (2.67) and (2.68) by the special irrational number, the golden mean! For example, according to (2.67) and (2.68) we can represent the Lucas number 3(n=2) and the Fibonacci number 5(n=5) as follows: 2 −2 1+ 5 1+ 5 + 3= 2 2 , (2.69) Chapter 2 Fibonacci and Lucas Numbers 93 5 −5 1+ 5 1+ 5 + 2 2 . (2.70) 5= 5 It is easily proven that the identity (2.69) is valid because according to (2.60) we have: 2 −2 1+ 5 1+ 5 L2 + F2 5 3 + 5 L−2 + F−2 5 3 − 5 = = = = and . 2 2 2 2 2 2 If now we substitute these expressions to the right of the formula (2.69), then the righthand side of (2.69) can be represented as follows: 3+ 5 3− 5 + . (2.71) 2 2 By adding the items of (2.71), we can see that all “irrationalities” in (2.71) are mutually eliminated and we obtain the number 3 from the sum, that is, the identity (2.69) is valid. We can be convinced in the validity of the identity (2.70), if we remember that according to (2.60) we have the following representations: 5 1+ 5 L5 + F5 5 11 + 5 5 = = 2 2 2 −5 1+ 5 L−5 + F−5 5 −11 + 5 5 = = and . 2 2 2 If we substitute these expressions into (2.70), then we obtain the follow ing expression for the righthand side of (2.70): 10 5 = 5, 2 5 whence follows the validity of the identity (2.70). Note that this reasoning has a general character, that is, for any Lucas or Fibonacci numbers that are given by formulas (2.67) and (2.68), all “irratio nals” in the righthand parts of (2.67) and (2.68) are always mutually elimi nated and we obtain integers as the outcome! 2.8.3. A Historical Analogy The situation of the mutual elimination of all “irrationals” in the formulas (2.67) and (2.68) reminds one of the situations that appeared in mathematics with the introduction of complex numbers. In the 16th century, Italian mathe maticians contributed significantly to the development of algebra: they solved THE MATHEMATICS OF HARMONY Alexey Stakhov 94 radicals with equations of the 3rd and 4th degrees. Cardano’s famous book, The Great Art (1545), contains the algebraic solution of the cubic equation: (2.72) x 3 + px + q = 0 according to the formula x = u + v, where 2 3 2 3 q q q p q p p + + ; v = 3 − − + ; uv = − . 2 2 2 3 3 2 3 It was proven algebraically that there are three roots of the algebraic equa tion (2.72), namely: 2 3 q p 1. For the case ∆ = + > 0 , Eq. (2.72) has one real root and 2 3 two complex conjugate roots; for example, the equation x3+15x+124=0 with ∆>0 has the following roots: x1=4; x2,3 = 2 ± 3i 3 . 2. For the case ∆=0, p≠0, q≠0, Eq. (2.72) has three real roots; for exam ple, the roots of the equation x312x+16=0 are: x1=4, x2,3=2. 3. For ∆<0 we have the most interesting instance, the socalled “non reducible” case, where we need to extract the root of the 3rd degree from complex numbers and the cubic roots u and v are complex numbers. Nev ertheless, in this case the equation (2.72) has real roots. For example, the equation x321x+20=0 with u=3 − (2.73) ∆ = −243, u = 3 −10 + −243 , v = 3 −10 − −243 has real roots 1, 4 and 5 ! This fact seemed paradoxical to the 16th century mathematicians! Real ly, all factors of the equation x321x+20=0 are real numbers, all its roots are real numbers, however, the intermediate calculations result in the “imagi nary,” “false,” “nonexistent” numbers of the kind (2.73). Mathematicians were in a very difficult situation again (starting with the discovery of irratio nals). To completely ignore the numbers of the kind (2.73) would mean to refuse the general formulas for the solution of algebraic equations of the 3rd degree, and other remarkable mathematical achievements. On the other hand, to recognize that these obtrusively appearing “monstrous” numbers such as (2.73) are equivalent with real numbers was inadmissible from the point of view of common sense. For a long time the “monstrous” numbers of the kind (2.73) were not recognized by many mathematicians. For example, Descartes considered that the complex numbers do not have any real interpretation and are doomed forever to remain only “imaginary” numbers (the name “imaginary numbers” came into mathematics in 17th century after Des cartes). Many eminent mathematicians, in particular, Newton and Leibniz, adhered to the same opinion. Chapter 2 Fibonacci and Lucas Numbers 95 17th Century English mathematician Vallis in the book Algebra: Histori cal and Practical Treatise (1685) pointed out a possible geometric interpreta tion of complex numbers. Ultimately, the complex numbers came into use in 18th century after the works of the French mathematician Moivre (16671754) who introduced the following wellknown formula: ( cos ϕ ± i sin ϕ ) n = cos nϕ ± sin nϕ. (2.74) After the introduction of Moivre’s formula (2.74), a representation of the complex numbers in trigonometric form came into use, which facilitated a solution of numerious mathematical problems. However, the famous Euler’s Formulas became the “moment of celebration” for complex numbers. By using Moivre’s formula (2.74), Euler proved the following formulas for trigonomet ric functions: e xi + e − xi e xi − e − xi , sin x = . (2.75) 2 2i Note that finding the connection between trigonometric and exponential functions expressed by Euler’s formulas emphasize a fundamental connection between the numbers π and е, two numerical constants of mathematics, that play an important role in mathematics, in particular, in the development of the complex number concept. By returning back to Binet formulas (2.67) and (2.68) and taking into consideration our reasoning concerning the complex numbers, we may sup pose that Binet formulas touch upon some rather deep numbertheoretical problems that are at the intersection of integers (Fibonacci and Lucas num bers) and irrationals (the golden mean). Further, in Chapter 9 we will try to broaden this idea by considering the number systems with irrational bases, which may overturn our ideas about number systems. cos x = 2.9. Fibonacci Rectangle and Fibonacci Spiral 2.9.1. Fibonacci Rectangles Let us consider the Fibonacci series: 1, 1, 2, 3, 5, 8, 13, 21, …. Take two squares with all of the sides equal to 1 (the area of each square is equal to 1) and put them together. As a result we get the 2×1 or “double square” rectan gle. Then, construct a new square of the size 2×2 on the longer side of the “double square.” Here we obtain a new rectangle of size 3×2. Then, construct Alexey Stakhov THE MATHEMATICS OF HARMONY 96 a new square of size 3×3 on the longer side of the preceding rectangle; as a result, we obtain a new rectangle by the size 5×3. Continuing this process, results in rectangles, with side equal to adjacent Fibonacci numbers, that is, the rectangles with the following sizes: 8×5, 13×8, 21×13, and so on (Fig. 2.4). 5 5 8 1 1 1 1 1 2 3 1 1 2 3 1 1 2 3 1 1 2 Figure 2.4. Fibonacci rectangles Such rectangles are called Fibonacci Rectangles. Consider a sequence of the ratios of the Fibonacci rectangles sides: 1:1, 2:1, 3:2, 5:3, 8:5, 13:8, 21:13, 34:21, …. As these ratios aim for the golden mean, we can conclude that the Fibonacci rectangles aim for the golden rectangle, that is, the Fibonacci rect angles are sequential approximations of the golden rectangle. 2.9.2. Fibonacci Spiral Fibonacci rectangles in Fig. 2.4 consist of the squares with sides 1, 2, 3, 5, 8, 13, …. Now, in each square we draw an arc that 34 is equal to a quarter of a circle as 55 shown in Fig. 2.5. If we connect 5 8 these arcs, we obtain a curve that 21 is suggestive of a spiral by its form 13 (Fig. 2.5). Strictly speaking, this curve is not a spiral from the Figure 2.5. Fibonacci spiral mathematical point of view. How ever, it is a very good approximation of the golden spirals that are widely met in nature. The curve in Fig. 2.5 we call a Fibonacci Spiral. 2.9.3. Fibonacci Spirals in Nature The great poet and scientist Goethe considered a spiral form as one of the characteristic attributes of all living organisms, as a manifestation of the most secret essence of life. Short moustaches of plants and horns of rams are twisted by a spiral, a growth of tissues in trunks of trees is carried out by a spiral, sun Chapter 2 Fibonacci and Lucas Numbers 97 flower seeds are disposed by a spiral, and so on. Everyone admires the forms of shells that are constructed under the spiral law. Let us begin from the spiral form of the Nautilus (Fig. 2.6a). If we compare its shape with the Fibonacci spiral (Fig. 2.5), we can conclude that the nautilus shell is constructed accord ing to the Fibonacci spiral principle. The motives of the Fibonacci spiral can be found in the shape of the Galaxy (Fig. 2.6b) and other sea shells (Fig. 2.6c). a) b) Figure 2.6. Fibonacci spirals in Nature c) 2.10. Chemistry by Fibonacci 2.10.1. Law of Multiple Ratios There is an opinion that the accumulated sum of knowledge in some disci pline can be called science only when it passes into a precise, quantitative analysis beyond the merely qualitative perspective. To the end of the 18th century, chemistry accumulated a significant volume of knowledge and chem ists had learned to decompose many complex substances into simple ones and to build the complex substances from the simple ones. During this process the problem of quantitatively describing various chemical compounds composing complex compounds appeared. It was necessary for the creation of a graceful theory of chemical structures, which would meet the requirements for the pro duction practice of various chemical products. The Law of Constant Proportions of Chemical Compounds is one of the funda mental chemical laws. This law came into chemical science after research by French scientist Prust (17541826). By studying chemical compounds, in particular the oxides of metals, he came to the conclusion that the chemical compounds have a strictly constant structure that is not dependent on the conditions of their forma Alexey Stakhov THE MATHEMATICS OF HARMONY 98 tion. Thanks to the works of the English scientist Dalton (17661844), the atom ic doctrine came into chemistry and the Law of Simple Multiple Ratios was formu lated. According to this law, simple integervalued ratios between atoms in chem ical structures exist. Now students know that the structure of water is described by the formula H2O, of common salt NaCl, and zinc oxide ZnO. Chemistry became a precise science. A new branch of chemistry arose that studies the ratios of atoms in chemical structures called Stoichiometry. The discovery of the law of multiple ratios is one of the remarkable achieve ments of chemistry as a science: a beautiful simplyordered system of chemical compounds appeared from the chaos of atomic representations. Atoms of differ ent elements can form different combinations that are connected by the forces of chemical connection. However, only some of them are stable; the other chemical combinations perish by disintegrating into more stable compounds. Only those combinations of atoms of different elements will be steady, if they correspond to simple integervalued ratios of their components. This idea is surprisingly simple and clear and corresponds completely to the Pythagorean doctrine about the dom inating role of integer numbers in the organization of the Universe! However, such formulation of the main chemical law evokes some bewil derment. It is not clear as to what the “simple integervalued ratios” of atoms in the formulas of chemical compounds really means. While we studied rather sim ple chemical compounds, atomic ratios in them usually corresponded to “small” numbers, for example, H2O, Al2O3, Fe3O4, As2O5. However, the range of the stud ied chemical compounds started to broaden rapidly. Formulas of chemical com pounds appeared there with the stoichiometric factors of 7, 9, 15, 21, etc. When the chemists started studying the structure of the organic compounds, it be came inconvenient to speak about the “simple integervalued ratios.” The chem ical structure of the bacteriophage is a peculiar champion in stoichiometry be cause it has the following formula: C5750H7227N2215O4131S590. What do the ratios of the “small” integers mean in this formula? Here we see fourdigit numbers! Thus, not all is so simple in stoichiometric laws: simplicity here is com bined with complexity, and the question about the possible ratios of atoms in compounds remains open. 2.10.2. Research by the Ukrainian Scientist Nikolai Vasyutinsky We will not delve too deeply into the chemistry of different compounds. We are interested in only one problem — whether Fibonacci numbers are found in the formulas of chemical compounds. The Ukrainian chemist Nikolai Vasyutinsky attempted to give the answer to this question in his book [31]. Chapter 2 Fibonacci and Lucas Numbers 99 By analyzing oxides of uranium and chromium, Vasyutinsky found that these chemical compounds are based on Fibonacci numbers. With the oxides of uranium, the structure of the generating oxides is changing not continu ously, but spasmodically, from one steady compound with a certain integer valued ratio of atoms into another one. Between the uranium oxides UO2 and UO3 a lot of the intermediate compounds are forming; their structures are described by the formulas U2O5, U3O8, U5O13, U8O21, U13O34. We can see that the atomic ratios in these compounds are equal to the ratios of Fibonacci num bers: 2:5, 3:8, 5:13, 8:21, 13:34. It is easy to prove that such ratios aim in the limit for the quadrate of the golden mean! Each of the described uranium ox ides can be represented as the sum of two oxides UO2 and UO3 taken in the different proportions, for example: U5O13=3UO3 +2UO2, U8O21=5UO3+3UO2. Here the factors of the oxides UO2 and UO3 correspond to the adjacent Fi bonacci numbers! This means that the structures of the aboveconsidered ura nium oxides are subordinated completely to the Fibonacci number regularity. Note that according to Vasyutinsky’s research the chromium oxides Cr2O5, Cr3O8, Cr5O13 have structures that are described by Fibonacci numbers. By considering the equalities of the type U5O13=3UO3 +2UO2, we can find their similarity to the algebraic golden mean equation of the 4th degree x4=3x+2 that is used for the description of the butadiene structure. By com paring the equality U8O21=5UO3+3UO2 with the algebraic golden mean equa tion of the 5th degree x5=5x+3, we can see that they also have a similar math ematical structure. Perhaps these analogies can be the beginning of rather interesting research into the field of stoichiometry. It is generally accepted procedure to determine the structures of the chemical compounds by the ratio of atoms of the elements that is included in this compound. However, it is possible to estimate the chemical com pounds by the ratios of the atoms (ions) of different elements to the mo bile valence electrons that are responsible for the formation of chemical connections between atoms. So, for example, in the chromium oxide Cr 2O5 the 10 valence electrons correspond to the 7 atoms of the chromium and oxygen. Making similar calculations for all the above oxides, we obtain the following ratios of the sums of the atoms to the sums of the valence electrons: 10:7, 16:11, 26:18, 42:29, 68:47. Note that the numerators of these fractions are double Fibonacci numbers (10=2×5, 16=2×8, 42=2×21, 68=2×34) and the denominators are the Lucas numbers 7, 11, 18, 29, 47. If we decreased sequentially the numerators and the denominators of these fractions on the Fibonacci numbers that correspond to the metal atom quantity in compounds, that is, on 2, 3, 5, 8, 13, we obtain the series of Alexey Stakhov THE MATHEMATICS OF HARMONY 100 ratios of adjacent Fibonacci numbers: 8:5, 13:8, 21:13, 34:21, 55:34; in the limit these ratios aim for the golden mean. Thus, the Ukrainian chemist Vasyutinsky convincingly demonstrated that chemical compounds organized by Fibonacci Numbers do exist. 2.11. Symmetry of Nature and the Nature of Symmetry 2.11.1. Basic Concepts of Symmetry Symmetry is one of the most fundamental scientific concepts that among the concept of Harmony has a relation to practically all branches of Nature, Science and Art. The outstanding mathematician Hermann Weil evaluated the role of symmetry in modern science in the following words: “Symmetry, as though wide or narrow we did not understand this word, there is an idea, with the help of which a person attempts to explain and cre ate an order, beauty and perfection.” What is symmetry? When we look at a mirror we can see us in its reflec tion; this is an example of mirror symmetry. The mirror reflection is an exam ple of the socalled Orthogonal Transformation that changes an orientation. In the general case Symmetry in mathematics is perceived as a transformation of space (plane), when each point M of space (plane) turns into the other point M′ with respect to some plane (or a straight line) a; here, the line segment MM is perpendicular to a plane (or a straight line) a and is divided by it in half. The plane (or the straight line) a is called Plane (or Axis) of Symmetry. Plane of Symmetry, Symmetry Axis and Center of Symmetry are fundamental concepts of symmetry. A plane of symmetry P is a plane that divides the figure into two mirrorsymmetrical parts that are disposed one to another as some subject and its mirror reflection. For example, the isosceles triangle ABC shown in Fig. 2.7 at B p A D C Figure 2.7. A symmetry of triangle and rectangular parallelepiped Chapter 2 Fibonacci and Lucas Numbers 101 the left is divided by the altitude BD into two mirrorsymmetrical halves ABD and BCD; thus, the altitude BD is the “track” of the plane of symmetry P that is perpen dicular to the plane of the triangle. We say that the isosceles triangle ABC has only the plane of symmetry P. In Fig. 2.7 at the right a rectangular parallelepiped (match box) is shown; it has three orthogonal planes of symmetry 3P. It is easy to show, that a cube has nine planes of symmetry — 9P. These examples could be continued. A symmetry axis L is called a straight line, around which a symmetrical figure can be turned several times in such a manner that each time the figure coincides with itself in space. For example, the equilateral triangle has the symmetry axis L3 that is perpendicular to the triangle plane in the center of the triangle. It is clear that there are three ways of turning a triangle around the symmetry axis L3 when the triangle coincides with itself. It is clear, that a square has the symmetry axis L4 and the pentagon L5. The cone also has a symmetry axis; there are an infinite number of the turns of the cone around the symmetry axis. This means that the cone has the symmetry axis of the type L∞. Finally, a symmetry center C of a figure is constructed inside the figure when any straight line, drawn through the point C, meets the identical points on the figure at equal distances from the center C. A sphere is a perfect exam ple of a figure with center symmetry. 2.11.2. Symmetry of Crystals For many centuries, the geometry of crystals seemed a mysterious and un solvable riddle. In 1619, the Great German mathematician and astronomer Jo hannes Kepler (15711630) paid close attention to the sixfold symmetry of snow flakes. He tried to explain it by the fact that their crystals are constructed from small identical balls that are connected one to another. Kepler’s idea was devel oped by the English polymath Robert Hooke (16351703) and the Russian poly math Mikhail Lomonosov (17111765), who made important contributions to Russian literature, education and science. They also assumed that it is possible to liken the elementary particles inside crystals to densely packed balls. Presently the Principle of Dense Spherical Packing underlies structural crystallography, only the spherical particles of the ancient authors are replaced now by atoms and ions. Fifty years after Kepler, the Danish geologist and crystallographer Nicolas Stenon (16381686) for the first time formulated the main idea of crystal growth: “Crystal growth is implemented not from within, similar to plants, but by means of the su perposition on the external planes of the crystal’s smallest particles, which are brought from outside by certain liquids.” This idea about crystal growth as the outcome of the sediment forming new stratums of substance on crystal faces preserves its sig Alexey Stakhov THE MATHEMATICS OF HARMONY 102 nificance until now. It was Nicolas Stenon who discovered the main law of geomet ric crystallography — the law of Constancy of Interface Angles. The law of Constancy of Interface Angles became a reliable basis for the de velopment of geometric crystallography and provided the richest material for ver ification of the true symmetry of crystallographic structures. The French researcher Rene Just Hauy assumed that the crystals consist not of the small balls (according to the assumption of Kepler, Hooke and Lomonosov), but of the molecules of parallelepiped shape, and these molecules are the extreme small splinters of the same shape. In other words, the crystals are peculiar “masonries” forming molec ular “bricks.” Despite all the naivety of this theory from the modern point of view, this idea did play an important role in the history of crystallography and gave an impetus to the origin of the theory of crystal lattice structures. The French crystallographer Brave was Hauy’s direct follower. He replaced the molecular “bricks” by points, the centers of molecular weight. As a result of this approach, the spatial lattices of crystals were obtained. By developing a hypothesis about the lattice structure of all crystals, Brave created the fun damentals of modern structural crystallography long before the experimental research of crystal structures with the help of Xrays. We can see in Fig. 2.8 the spatial crystal’s lattices of the common salt NaCl (a) and the calcium oxide CaO (b). In all crystals, we can see the set of the identical atoms located like the points of a spatial lattice. The straight lines, on which atoms in the lattice are located, are named rows, and the planes, filled by atoms, are named planar lattices. A presentation about the lattice structure of crystals allowed us to give a general definition of crystals. It is accepted that crystals are solid bod ies, which consist of particles (atoms, ions, molecules), located the strictly a) regular knotlike spatial lattice. This b) general definition can be applied to Figure 2.8. Spatial crystal lattices of NaCl (а) and CaO (b) all types of Nature’s crystals. However, already in 1830, long before Brave, the German professor Johannes Fridrich Hessel (17961872) published the long article Crystallometry, where he developed an approach to crystal geometry based on the symmetry concept. In this article, Hessel gave a full research report of the set of symmetry elements of crystals and spatial geometric figures. Only in 1890, that is, in 60 years after pub lication of this article, and in 18 years after the death of its author, crystallogra phers could evaluate the significance of Hessel’s outstanding discovery. Chapter 2 Fibonacci and Lucas Numbers 103 In 1890, the eminent Russian crystallographer Fedorov proved the exist ence of the 230 sets of symmetry elements for all existing types of crystals. Fedorov’s symmetry groups corresponding to the geometric laws of atom lo cation in crystal structure underlies modern structural crystallography. A presentation about the lattice structure of crystals, in particular, a con cept of “planar lattice,” on which the knots of the crystal lattice are located, resulted in the discovery of the Major Law of Crystallographic Symmetry. Ac cording to this law of crystal, symmetry axes the first, second, third, fourth and sixth orders are allowed. However, it is impossible for crystals to have the fifth order symmetry or anything greater than the sixth order of symmetry. According to the main crystallography law, there is a fundamental difference between the symmetry of the mineral world and the symmetry of the living world, where fivefold symmetry is widely used. For crystals, the fivefold axis of sym metry and the symmetry axes greater than the 6th order are prohibited: this strict rule of classical crystallography existed until 1982, when the Israeli physicist Dan Shechtman proclaimed the startling discovery of quasicrystals. 2.11.3. Symmetry Laws in Nature The concept of symmetry is used widely in physics. If the laws that determine relations be tween physical magnitudes and a change of these magnitudes in the course of time do not vary at the definite operations (transformations), they say, that these laws have symmetry (or they are invariant) with respect to the given transforma tions. For example, the law of gravitation is val id for any points of space, that is, this law is in variant with respect to the system of coordinates. In the opinion of outstanding Russian sci entist and academician Vernadsky, “symmetry encompasses properties of all fields of physics and chemistry.” The Pythagoreans paid very close attention to the phenomenon of symmetry in Nature that was connected with their Harmony doctrine. The two kinds of symmetry, Mirror and Radial, are widespread throughout Nature. A butterfly, a leaf, and a beetle (Fig. 2.9a) have “mirror” symmetry a) b) Figure 2.9. Natural forms with “bilateral” (а) and “radial” (b) symmetries Alexey Stakhov THE MATHEMATICS OF HARMONY 104 and often such type of symmetry is called Leaf Symmetry or Bilateral Symmetry. A mushroom, a chamomile, and a pine tree (Fig. 2.9b) have a “radial” symmetry, and often such type of symmetry is called a ChamomileMushroom Symmetry. Already in the 19th century the researchers in this area came to the con clusion that symmetry of natural forms largely depend upon the influence of the Earth’s gravitational forces that have the symmetry of a cone in each point. In the outcome, the following law was found: “Everything that grows or moves in a vertical direction, that is, upwards or down relative to the Earth’s surface is subordinated to the “radial” (“chamomile mushroom”) symmetry. Everything that grows and moves horizontally or with an inclination relative to the Earth’s surface is subordinated to the “bilateral” or “leaf” symmetry.” 2.11.4. Symmetry Laws in Art The principle of symmetry is used widely in Art. The curbs in architectural and sculptural works, the ornamental designs in the applied art are examples of the application of the laws of symmetry. Symmetry together with the golden mean principle is often used in works of art. Raphael’s picture The Engagement of Virgin Mary (Fig. 2.10) is just such an example. Figure 2.10. Raphael’s picture “The Engagement of Virgin Mary” and its harmonic analysis Chapter 2 Fibonacci and Lucas Numbers 105 2.12. Omnipresent Phyllotaxis 2.12.1. A Helical Symmetry Everything in Nature appears subordinate to stringent mathematical laws. For example, leaf disposition on plant stems also has stringent mathematical reg ularity, a phenomenon called Phyllotaxis in botany. The essence of phyllotaxis consists of a spiral disposition of leaves on plant stems, trees, petals in flower baskets, seeds in a pi necone and those of a sunflower head, etc. This phenomenon was already known to Kepler and was a subject of discussion of many sci entists throughout the years, in cluding Leonardo da Vinci, Turing, Figure 2.11. A helical symmetry Veil, and so on. Much more com posite concepts of symmetry are used in the phenomenon of phyllotaxis, in partic ular, the concept of Helical Symmetry. Let us examine, for example, a disposition of leaves on the plant stem (Fig. 2.11). We can see in Fig. 2.11 that the leaves are on different heights on the stem along the he lical curve that encircles the stem. To move up from the lower leaf to the higher one, it is 5 6 necessary virtually to turn the leaf at some angle around the vertical axis and then to raise 5 the leaf up a definite distance. This transfor 4 mation is the essence of helical symmetry. 4 Let us examine typical helical axes that can appear along the plant stems (Fig. 2.12). 3 In Fig. 2.12a we can see the stem of a plant 3 with helical symmetry axis of the 3rd order. We can unite all leaves of the plant by any 2 2 virtual line. We start from leaf 1. To move 1 up from leaf 1 to leaf 2, it is necessary to turn 1 leaf 1 around the stem axis 120° counter clockwise (if looking from below) and then (a) (b) move up leaf 1 along the stem in the verti Figure 2.12. The helical axis on stems of plants cal direction until it coincides with leaf 2. If Alexey Stakhov THE MATHEMATICS OF HARMONY 106 we repeat a similar operation, we can move up from leaf 2 to leaf 3 and then on to leaf 4. It is necessary to note that leaf 4 is on the same line with respect to leaf 1 (as though it repeats on the higher level). It is clear if we want to move up from leaf 1 to leaf 4 we have to make three turns of 120o i.e. we have to carry out a full rotation around the stem axis (120°×3=360°). Botanists called the rotation angle of the helical axis the Leaf’s Diver gence Angle. A vertical straight line that is parallel to the stem of the plant connecting two leaves is called an Ortholine. The line segment of leaf 1 to leaf 4 of the ortholine corresponds to a full transformation of the helical axis. We will see further that the number of turns around the stem axis to move up from the bottom leaf to the next leaf directly above it along the ortholine, can be equal not only to 1, but also to 2, 3, and so. This number of turns is called the Leaf’s Cycle. In botany, it is acceptable to characterize helical leaf location by the ratio m/n where the numerator m is a number of turns in the leaf’s cycle, and the denominator n is the number of leaves in this cycle. In the above case, we have a helical axis of 1/3. In Fig. 2.12b the helical axis of fivefold symmetry with a leaf cycle 2 is presented; this means that for the transition from leaf 1 to leaf 6 it is necessary to make two full turns. The fraction 2/5 characterizes the given helical axis; the leaf divergence angle is equal to 144° that is, we have: 360/5=72°; 72°×2=144°. Note that there are also more complex helical axes, for example, of the kind 3/8, 5/13, and so forth. There is a question as to what values can take the numbers m and n that describe the helical axis of the kind m/n. And here Nature gives us the follow ing surprise called the Law of Phyllotaxis. Botanists assert that the fractions that describe the helical axes of plants build up a strict mathematical sequence of the following kind, for example: 1/2, 1/3, 2/5, 3/8, 5/13, 8/21, 13/34, …. (2.76) Note that the fractions in the sequence (2.76) are the ratios of two adja cent Fibonacci numbers beginning with number one. It is easy to prove, that the sequence (2.76) aims for the number τ2=0.382, that is, to the inverse of the square of the golden mean. Botanists found that different plants have different ratios of the above sequence (2.76). For example, the fraction 1/2 is peculiar to cereals, birch and grapes; 1/3 to sedge, tulip, and alder; 2/5 to pear, currant, and plum; 3/8 to cabbage, radish, and flax; 5/13 to firtree and jasmine, etc. What is the “physical” cause that underlies the phyllotaxis law (2.76)? The answer is very simple. It proves to be the disposition of leaves that allow for the maximum inflow of solar energy to the plant. Chapter 2 Fibonacci and Lucas Numbers 107 2.12.2. Densely Packed Phyllotaxis Structures The phyllotaxis phenomenon shows itself in inflorescences and densely packed botanical structures, such as, pinecones, pineapples, cacti, sunflowers, cauliflowers and many other structures. Figure 2.13 gives examples of phyllotaxis objects (cactus, sunflower, cone flower, Romanescue cauliflower, pineapple, pinecone), in which the phyllo taxis law is based on the numerical sequence (2.30) that consists of the ratios of adjacent Fibonacci numbers. This means that the seeds or small parts on the surface of such botanical objects are located at the crossings of lefthand and righthand spirals; here the ratio of the numbers of lefthand and right hand spirals is always equal to the ratio of adjacent Fibonacci numbers (2.30). The same regularity is observed in baskets of flowers. (a) (b) (c) (d) (e) (f) Figure 2.13. Phyllotaxis structures: (a) cactus; (b) head of sunflower; (c) coneflower; (d) Romanescue cauliflower; (e) pineapple; (f) pine cone Geometric models of phyllotaxis structures in Fig. 2.14 provide a clearer representation of this unique botanical phenomenon. Alexey Stakhov THE MATHEMATICS OF HARMONY 108 (a) (b) (c) Figure 2.14. Geometric models of phyllotaxis structures: (a) pineapple; (b) pinecone; (c) shasta daisy Thus, we find strict mathematics in the dispositions of leaves on plant stalks, flower petals, in the crosssection of an apple (pentacle), in the spiral dispositions of seeds of a pinecone, pineapple, cactus and the head of a sun flower. And this mathematical law is expressed by Fibonacci numbers and therefore the golden mean! Once again, it appears as though Nature is subor Chapter 2 Fibonacci and Lucas Numbers 109 dinate to a coherent plan, the uniform Law of the Golden Section! To discover and explain this fundamental law of Nature in all its expressions is one of the most important problems of science and philosophy. 2.12.3. Use of Phyllotaxis Lattices in Painting In the above we mentioned a wideranging use of the golden rectangle and other golden geometrical figures in artwork. One art method used by artists of the Renaissance was the use of phyllotaxis grids in painting. This is not unlike the botanical phyllotaxis phenomenon in which Nature appears to design pinecones, pineapples, cacti, sunflower heads, and many other botanical structures. According to phyllotaxis law, cactus spines are located along Fibonacci spirals; here the adjacent Fibonacci numbers 21 and 34 are the numbers of the lefthand and righthand spirals. If we unroll on a plane the spines of a cactus, we can obtain a raster grid (Fig. 2.15). In the raster grid in Fig. 2.15 the in clined lines with righthand and lefthand inclination represent the principle of disposition of the spines on the cactus surface. This raster grid has 21 lines with righthand inclination and 34 lines with lefthand inclination. The net of the lines in Fig. 2.15, Phyllotaxis Raster Grid, from an aesthetic point of view looks as if it is a golden rectangle. Many artists of the Renaissance used the phyllotaxis raster grid in their art work. Figure 2.15. Phyllotaxis raster grid Paturi, the Austrian researcher and author of the remarkable book, Plants as Ingenious Engineers of Nature [140], made an analysis of the use of phyllotaxis raster grids in the artworks of great artists. For this purpose, he put together the phyllotaxis raster grid with Titian’s picture Bacchus and Ariadne (Fig. 2.16). After an analysis of Titian’s painting Paturi concluded: “All basic lines of the perspective coincide with the raster grid. Even the set of the minor parts and forms were placed by the artist in that field of internal constraints, on which the picture is constructed. Pay attention to the line along the small hill seen on the horizon near to the church campa nile, to the branches of the large tree, to the outline of the cloud lying under the constellation, to the hind paws and the belly line of the large wild cat, Alexey Stakhov THE MATHEMATICS OF HARMONY 110 Figure 2.16. Titian’s Bacchus and Ariadne along the direction of the axis of the over turned vase, to the raised right hand of the satyr in the garland of grapevines in the right corner of the canvas and, at last, to the raised leg of the horse.” Paturi further concluded: “At all times the artists consciously or un consciously comprehended the laws of aesthet ic perception by watching nature. Artists were always enchanted by the simple and simulta neously rational geometry of the growth of bi ological forms.” 2.13. “Fibonacci Resonances” of the Genetic Code 2.13.1. The Initial Data about the Genetic Code Among the biological concepts [141] that are well formalized and have a level of general scientific significance, the genetic code takes special prece dence. Discovery of the striking simplicity of the basic principles of the genet ic code places it amongst the major modern discoveries of mankind. This sim plicity consists of the fact that inheritable information is encoded in the texts from threelettered words — triplets or codonums compounded on the basis of the alphabet that consists of the four characters or nitrogen bases: A (Ade nine), C (Cytosine), G (Guanine), T (Thiamine). The given system of the ge netic information represents a unique and boundless set of diverse living or ganisms and is called the Genetic Code. 2.13.2. DNA SUPRAcode (JeanClaude Perez’s Discovery) In 1990 JeanClaude Perez, an employee of IBM, made a rather unexpected discovery in the field of the genetic code. He discovered the mathematical law that controls the selforganization of bases A, C, G and Т inside the DNA. He found that the consecutive sets of the DNA nucleotides are organized in frames of remote order called “RESONANCES.” Here, the resonance means a special proportion that divides the DNA sequence according to Fibonacci numbers (1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144 …). Chapter 2 Fibonacci and Lucas Numbers 111 The key idea of Perez’s discovery, called the DNA SUPRAcode, consists of the following. Let us consider some fragment of the genetic code that consists of the A, C, G and T bases. Suppose that the length of this fragment is equal to some Fibonacci number, for example, 144. If a number of the Tbases in the DNA fragment is equal to 55 (Fibonacci number), and a total number of the C, A and G bases is equal to 89 (Fibonacci number), then this fragment of the genetic code forms is a resonance — that is, a proportion between three adjacent Fibonacci numbers (55:89:144). Here it is permissible to consider any combina tions of the bases, that is, C against AGT, A against TCG, or G against TCA. The discovery consists of the fact that the arbitrary DNAchain forms some set of the resonances. As a rule, the fragments of the genetic code of the length equal to the Fibonacci number Fn are divided into the subset of the Tbases, and the subset of the remaining A, C, G bases; here the number of Tbases is equal to the Fibonacci number Fn2 and the total number of the remaining A, C, G bases is equal to the Fibonacci number Fn1, where Fn=Fn1+Fn2. If we make a systematic study of all the Fibonacci fragments of the genetic code, we can obtain a set of the resonances that is called the SUPRAcode of DNA. 2.13.3. A Verification of JeanClaude Perez’s Law In the Petoukhov book [141] the sequences of triplets for the α and β chains of insulin are given. For the βchain this sequence has the following form: ATGTTGGTCAATCAGCACCTTTGTGGTTCTCACCTCGTT GAAGCTTTGTACCTTGTTTGCGGTGAACGTGGTTTCTTC TACACTCCTAAGACT Note that all the Tbases in the indicated sequence are marked by bold type. A verification of JeanClaude Perez’s Law using the βchain of the insulin molecule as an example (see above) results in the following outcome. The total number of the triplets in the βchain is 30, that is, the molecule contains 90 bases (the nearest Fibonacci number is 89). If we count the number of Tbases in the above βchain, we find that there are 34 (a Fibonacci number). Then the number of the remaining A, C, G bases is equal to 9034=56 (the nearest Fibonacci num ber being 55). Thus, there is the following proportion between the Tbases and the rest, i.e. the A, C, G bases in the βchain are 90:56:34. This proportion is very close to the Fibonacci resonance of 89:55:34. It follows from this analysis that JeanClaude Perez’s Law for the insulin βchain is fulfilled with great accuracy. If now we take the initial segment of the above βchain that consists of the first 18 triplets, that is, of the 54 bases (the nearest Fibonacci number is 55), and count the number of Tbases in this fragment we find that it is 22 (the nearest Fibonacci Alexey Stakhov THE MATHEMATICS OF HARMONY 112 number being 21). This means that we have the following proportion in the first fragment of the βchain, namely, 54:32:22 that is close to a Fibonacci resonance of 54:34:21.Thus, the Perez’s Law is also fulfilled for the first fragment. If we take the segment that consists of the rest of the 12 triplets (36 bases) then the number of the Tbases in this segment is 12 (the nearest Fibonacci number being 13). Thus, for this case we have a proportion of 34:24:12 that is close to the Fibonacci reso nance of 34:21:13. Therefore, both for the βchain of the insulin molecule as a whole, and for its separate fragments, the Perez Law is fulfilled in practice with sufficient accuracy. In addition, it is possible to see that practically in any segment of the βchain there is a tendency towards the golden mean. This surprising discovery of JeanClaude Perez has a number of interest ing applications in the socalled plastic arts and market analysis. Below we can see application of this law in, for example, music, poetry, cinema and mar ket processes (Elliott Waves). 2.14. The Golden Section and Fibonacci Numbers in Music and Cinema 2.14.1. Pythagorean Theory of Musical Harmony Music is a form of art reflecting reality and influencing a person by means of organized sound sequences. Preserving some similarity to natural sounds, musical sounds principally differ from the latter by constrained pitch and rhythmic organiza tion (musical harmony). Since antiquity, the search for the “laws of musical harmony” has been one of the important pursuits of science. Greek musicology does not corre spond exactly with the character of modern musicology. Greek musicology was not directed toward analyzing musical works. Rather, the Greeks saw musicology as a study of the acoustic aspects of sound. An essential feature of Greek musicology was the aspiration for a mathematical description of musical harmony. Pythagoras is credited with the discovery of two fundamental harmonic musical laws: 1. If the ratio of the oscillation frequencies of two tones is describable in small integers, a harmonic sound may result; 2. To get a harmonic triad, it is necessary to add a third tone to the chord, which consists of two consonant tones. The oscillation frequency of the third tone should be in harmonic proportional connection to the two tones of the chord. Pythagoras’ work on the scientific explanation of musical harmony was the first Chapter 2 Fibonacci and Lucas Numbers 113 wellgrounded scientific theory of musical harmony! Pythagoras attempted to apply his musical theory to cosmology; according to his ideas, the planets of the Solar Sys tem are arranged pursuant to the musical octave (“harmony of the spheres”). There are many legends surrounding the acoustic experiments of Pythag oras. The most popular of them describes how Pythagoras, passing by a black smith, heard the sounds of hammer blows on an anvil and found in them an octave, a fifth and a fourth. In his excitement, Pythagoras hastened to the smithy and after a series of experiments with hammers, found that the differ ence in sounds depended on the weight of the hammers. The Pythagorean discovery of the deep connection between music and mathematics caused harmonics to be included amongst the mathematical sci ences influencing the future development of musical theory. 2.14.2. Chopin’s Etudes in the Lighting of the Golden Section Any piece of music has a temporal duration and is divided into separate parts by marks that attract our attention (“expression marks”) and simplify our perception of the musical work. These marks are dynamic culmination points within the musical work. The question is whether or not there is cer tain regularity in the placing of “expression marks” in the musical work. An attempt to answer this question was made by Russian musicologist Sabaneev. He showed [142] that the separate time intervals of the musical works of Chopin are connected, as a rule, by “culmination events” dividing the musical work in the golden ratio. Sabaneev writes: “All such events are referred by the author’s instinct to such points of the musical work that divide the temporal durations into separate parts being in the ratio of the “Golden Section.” Observation shows that quite often similar “expression marks” are found at golden points. It is quite surprising, because frequently the knowledge about these things is absent for many composers, and these facts are a consequence of internal feelings of rhythm.” Sabaneev’s analysis of a large number of musical works allowed him to con clude that the organization of music is often done so that its cardinal parts, divided by the “marks,” are based upon the golden mean series. Such organization of music corresponds to the most economical perception and consequently creates an im pression of the best “regularity” within the musical structure. In Sabaneev’s opin ion, quantity and frequency of golden mean appearances in music depends upon a “composer’s expertise.” The musical work of great composers is distinguished by the highest percentage of golden mean appearances, that is, “the intuition of the form and regularity, as would be expected, is highest for the first class geniuses.” Alexey Stakhov THE MATHEMATICS OF HARMONY 114 According to Sabaneev, the golden mean is repeatedly met in the musical works of different composers. Every such golden point reflects the overall mu sical event, the qualitative jump in the development of the musical theme. In the 1,770 musical works of 42 composers that were studied by him, the golden mean is met 3,275 times. The greatest frequency of musical works based upon the golden mean is found in Beethoven (97%), Haydn (97%), Arensky (95 %), Chopin (92%), Mozart (91%), Schubert (91%), and Scriabin (90 %). Chopin’s 27 etudes were studied by Sabaneev in particular detail. Golden sections were found 154 times in them; the golden section being absent in only three etudes. In some cases, the principles of symmetry and the golden mean are combined simultaneously in the structure of a musical work; in these cases, the musical work is divided into some symmetrical parts; however, the golden section is met within every separate part. For example, Beethoven’s music is often divided into two symmetrical parts; however, the golden sec tion is always met inside each part. 2.14.3. Rosenov’s Research The Russian art critic Rosenov paid particular attention to research into the harmonic laws of music. He asserted that the stringent proportional rela tions in music and poetry that are present with the golden mean should play an outstanding role. Rosenov chose for his analysis a number of typical musical works of out standing composers: Bach, Beethoven, Chopin and Wagner. For example, by studying Bach’s Chromatic fantasy and fugue, he used the quarter duration as the unit of measure. There are 330 such units of measure in this musical work. The 204th quarter from the beginning corresponds to the golden section here. This golden point coincides precisely with the fermata (in musical pieces the fermata sign augments a pause’s duration, usually in 1.52 times). Further more, the golden point separates the first part of the musical work (the pre lude) from the second part. The fugue, which follows the fantasy, also demon strates an astonishing harmony of parts. Looking at the scheme of the fugue’s harmonic analysis, “one experiences a sacred thrill [and] comes into contact with the greatness of the composer, who embodies the innermost laws of cre ative musical work with such accuracy.” Rosenov analyzed the final of the sonata of Cismoll by Beethoven, the FantasiaImpromptu by Chopin, the introduction to Tristan and Isolda by Wagner and so on. The golden section occurs very often in all of these musical works. He paid special attention to Chopin’s Fantasia, which was an impromptu Chapter 2 Fibonacci and Lucas Numbers 115 creation and not subject to any editing. Since this was impromptu, there was no conscious application of the golden section, which nevertheless is present throughout this musical piece even in the smallest musical parts. Thus, we should recognize the golden section as a harmonic criterion in many musical works appearing frequently. Every part can be divided into two parts by the golden section and each new part can be divided by the golden section. Here we can see an analogy with Fibonacci resonances of the genetic code where each segment of the genetic code is divided in the golden ratio! 2.14.4. Bella Bartok’s Music The Hungarian Bella Bartok is one of the composers who consciously used the golden section in his music. The deep harmonic analysis of Bartok’s music was giv en by Jay Kappraff [47] who wrote: “Bartok based his music on the deepest layer of folk music. He believed that all folk music of the world can ultimately be traced to a few primeval sources.” According to the Hungarian musicologist Erno Lendai, “[Bartok] discovered and drew into his art the laws governing the depths of the human soul which have been untouched by civilization.” As Kappraff noted, “Bar tok based the entire structure of his music on the golden mean and Fibonacci series — from the largest elements of the whole piece, whether symphony or sonata, to the movement, principal, and secondary themes and down to the smallest phrase.” 2.14.5. The Golden Mean and Fibonacci Numbers in Cinema Sergey Eisenstein (18981948) is the great Soviet film director and art theo rist. Having seen inexhaustible new opportunities for art in cinema, Eisenstein soon created his first film, Strike. In 1925 Eisenstein created the film The Battleship Potemkin that caused a triumph in the world. The American Film Academy recog nized The Battleship Potemkin as the best film of 1926. At the Parisian Exhibition of Arts this film received the highest award: “Super GrandPrix.” It became a classic of new cinema art. Creating the films October, Old and New, etc., Eisenstein developed a theory of “intellectual cinema” that gave the spectator an opportunity to get not only the artistic, but also scientific knowledge and concepts. Eisenstein was a pioneer in the use of the golden mean in cinema. It is most interesting how that Eisenstein used the “Golden Section Principle” in his film The Battleship Potemkin. First of all this film consists of 5 acts. The partitioning of the film into two separate parts corresponds to the proportion 2:3, that is, the first Fi bonacci approximation to the golden mean. This 2:3 watershed occurs between the end of the second and the beginning of the third act of the fiveact film, the basic caesura of the film occurs as a Zero Point of the action stop. However, perhaps, the Alexey Stakhov THE MATHEMATICS OF HARMONY 116 most curious fact in all of this is that the “Golden Section Law” in The Battleship Potemkin is observed not only for the zero point of the movement, but also is true for the apogee point: a Red Flag on the battleship mast. Finally, the moment of the appearance of the red flag is the golden section of the film’s duration. 2.15. The Music of Poetry In many respects, poetic works are close to music. A distinctive rhythm, regular alternation of stressed and unstressed syllables, a certain dimension ality, and their emotional saturation make poetry a native sister to music. Each verse has its musical form, rhythm and melody. It is possible to expect some musical features in poems, for example, regularity of musical harmony, and consequently, the golden section can often be present. 2.15.1. Pushkin’s Poetry Alexander Pushkin was a Romantic author consid ered to be the greatest Russian poet and the founder of modern Russian literature. Pushkin’s father came from a distinguished family of the Russian nobility. His moth er’s grandfather was Abram Petrovich Gannibal, an Ethi opian, who as a child was abducted by the Turks during their governing of Ethiopia, and he became a great mili tary leader, engineer and nobleman in Russia under the Alexander Pushkin auspices of his adoptive father, Peter the Great. Born in Moscow, Pushkin published his first poem at the age of fifteen. By the time he finished the first grade of the prestigious Imperial Lyceum in Tsar skoe Selo near St. Petersburg, his talent was widely recognized by the Russian literary community. In 1820 he published his first epic poem, Ruslan and Lyud mila. Because of his 1820 poem, Ode to Liberty, he was exiled by the Czar Alex ander I to the south of Russia. First coming to Kishinev in 1820, he there be came a Freemason. He was in Kishinev until 1823. After a summer trip to the Caucasus and to the Crimea, he wrote two Romantic poems which brought him wide popularity, The Captive of the Caucasus and The Fountain of Bakhchisaray. When Alexander’s brother, Nicholas I, came to power in 1825, he invited Push kin back to the capital and gave him a government post. However, Nicholas acted as his personal censor to make sure that Pushkin did not publish anything that Chapter 2 Fibonacci and Lucas Numbers 117 would hurt the government. The Czar ordered his spies to follow him, cutting out whole stanzas from his manuscripts. In the autumn of 1830 Pushkin left the cap ital to visit a small village left to him by his father. There Pushkin wrote some of his best poems, including completing his most famous poem Eugene Onegin. On January 19, 1831, when he was almost 30 years old, Pushkin married the beautiful young Nathalie Goncharova. Although they had three children, they were not a happy couple. Nathalie was very beautiful and a favorite at the court, often encouraging the attention of other men. In 1837, influenced by rumors that his wife had entered into a scandalous liaison, Pushkin chal lenged her alleged lover, Georges d’Anthиs, to a duel. Pushkin was mortally wounded in the duel and died on January 29, 1837. Pushkin’s poetic works were analyzed by many researchers from the gold en mean point of view. Let us begin with the poem’s size, the number of lines. At first, it seemed that this parameter of the poem could change somewhat arbitrarily. However, this proved not to be so. For example, the analysis of Pushkin’s poems by Vasjutinsky [31] showed that the sizes of Pushkin’s po ems are not distributed uniformly; Pushkin appearantly preferred poems of 5, 8, 13, 21 and 34 lines (Fibonacci numbers!). It was noted by many researchers that his poetical verses and poems are sim ilar to musical pieces; there are also culmination points in them dividing the po ems at the golden section. By studying Pushkin’s poem Shoemaker, Vasjutinsky noted that the poem consists of 13 lines. Two semantic parts can be singled out in the verse: the first part consists of 8 lines and the second one (the parable moral) consists of 5 lines. Note of course that the numbers 13, 8, 5 are Fibonacci numbers and their ratios are thus an approximation to the golden mean proportion! One of Pushkin’s last poems, Not Dearly I Appreciate the HighSounding Rights …, consists of 21 lines in which two semantic parts are singled out: the first part is 13 lines and the second part 8 lines (21, 13 and 8 are of course Fibonacci numbers). The analysis of the famous novel, Eugene Onegin, by Vasyutinsky [47] is very intriguing. This novel consists of 8 chapters; each chapter is, on the average, about 50 verses. The eighth chapter is the most perfect and the most emotional. This chapter consists of 51 verses. Together with the letter of Eugene to Tatjana, the size of this chapter corresponds precisely to the Fibonacci number 55. Vasyutinsky writes: “Eugeny’s declaration of love to Tatjana is the chap ter’s culmination (the line ‘To turn pale and to die away… it is bliss!” This line divides the 8th chapter into two parts, the first part consists of 477 lines and the second part consists of 295 lines. Their ratio is equal to 1.617! We can see the finest conformity to the golden mean! This fact is the great miracle of harmony created by Pushkin’s genius!” Alexey Stakhov THE MATHEMATICS OF HARMONY 118 2.15.2. Lermontov’s Poetry Rosenov analyzed many poetic works of Lermontov, Schiller and Tolstoy, also finding in them the golden mean. Michael Lermontov was born on the 15th of Octo ber, 1814 into a family of nobility. The poet spent his youth at Tarkhany. He enrolled in Moscow University, but very soon had to leave the University. Later he en rolled in and graduated from St. Petersburg School of Cavalry Cadets. In 1837, he was exiled to the Cauca sus because of his poem on Pushkin’s death, in which he blamed the ruling circles of Russia and the Czar Nico las I. In 1841, Lermontov was sent into exile in the Cau Michael Lermontov casus once again. As a result of intrigues between the officers, he was provoked into a personal quarrel with an old schoolfellow. This led him to a duel on July 15, 1841 in which the poet was killed. Lermontov began writing his poems when he was very young. However, he became famous because of his poem on Pushkin’s death. Lermontov’s poems, including The Demon and Mtsyri, and his innumerable lyrical poems such as, A Hero of Our Time and Masquerade, are masterpieces of Russian literature. Lermontov’s famous poem, Borodino (see English translation by Eugene M. Kayden at www.lermontov.net/content/view/39/2/), is divided into two parts: the short introduction and the main part. The introduction consists of only one stanza: “But tell me, uncle, why our men Let Moscow burn, yet fought again To drive the French away? I hear it was a dreadful fight, A bitter war, by day and night; That’s why we celebrate the might Of Borodino today.” The main part of the poem consists of 13 verses, and it is divided into two parts: first consist of 8 verses, and second of 5 verses. In the first part, a battle of Borodino is described with rising tension an expectation of action; in the second part, the battle itself is described with gradual reduction of tension to the end of the poem. Each verse consists of 7 lines, that is, in total the main part consists of 91 lines. If we divide the main part in the golden section (91/1.618 = 56.238), we find, that the division point corresponds to the beginning of the 57th line, that begins from the short phrase: “O what a Chapter 2 Fibonacci and Lucas Numbers 119 day!” This phrase represents the culmination point of the tense anticipa tion, which completes the first part of the poem the expectation of the action, and opens its second part a description of the battle: O what a day! The Frenchmen came, A solid mass, like clouds aflame, Straight for our redoubt. Their lancers rode with pennons bright; Dragoons came on in all their might Against our walls, and in the fight They scattered in a rout. 2.15.3. Shota Rustaveli’s Poem Shota Rustaveli was a Georgian poet of the 12th century. He is one of the greatest representatives of medieval literature. There is not much biographical data about Rustaveli: the ex act dates of his birth and death, and historical data about the main events of his life are unknown. Many poetic works writ ten by him were lost. He is the author of the literary work The Knight in the Panther’s Skin, a Georgian national epic poem. Rustaveli was a Georgian noble and treasurer to the Queen of Georgia Tamara. He also restored and painted frescoes in the Shota Rustaveli Georgian monastery of the Holy Cross in Jerusalem. One of the pillars of this monastery bears a portrait, which is believed to be that of the poet. According to the legend, he fell hopelessly in love with Tsarina Tamara. It was here in the cell of this monastery that Rustaveli terminated his life. A grandiose poem, The Knight in the Panther’s Skin, brought to Rustaveli world glory. Translated into many languages this poem is rightfully consid ered one of the world’s greatest pieces of literature. Many researchers of Shota Rustaveli’s poem note its harmony and melody. In the opinion of Georgian scientist and academician Zereteli, these properties of Rusataveli’s poem arise, thanks to his conscientious use of the golden mean in both its poetic forms and the construction of its verses. Rustaveli’s poem con sists of 1 637 stanzas; each consisting of four lines. Each line consists of 16 sylla bles divided into two equal parts, the halflines, which therefore consist of 8 syllables. All halflines are divided into two segments of two kinds: the Akind segment is the halfline with equal segments and with an even number of sylla bles (4+4), the Bkind segment is the halfline with asymmetrical division into two unequal parts (5+3 or 3+5). Thus, we can see the following 3:5:8 ratio in Alexey Stakhov THE MATHEMATICS OF HARMONY 120 the halfline of the Bkind; that is, Fibonacci numbers, whose ratios are, of course, close to the golden mean. Zereteli proved that 863 of the 1 637 stanzas in Rustave li’s poem are constructed according to the golden mean principle! 2.16. The Problem of Choice: Will Buridan’s Donkey Die? The Russian academician Zeldovich wrote in one of his works: “Do not treat with contempt “simple” reasoning. Those researchers, who obtain deep results out of the simple, but firmly established facts, deserve the highest praise.” The works of American psychologist Vladimir Lefevre have an utter “simplicity” of initial reasoning giving them aesthetic charm. 2.16.1. A Parable about Buridan’s Donkey Often times we have to choose the best amongst various possibilities. This leads to a rather unexpected consequence — the pleasure of choice suddenly turns into a complex problem, and this problem can be extremely serious. This problem of choice is a perennial problem. The Greek philosopher Aristotle wrote of the difficulty for a person experiencing both hunger and thirst who is equally distant from food and drink. He remains unmoved at the same place because he cannot make a choice. This problem of a person not being able to make a choice between food and drink, was formulated more precisely by the French philosopher Jean Buridan, who demonstrated an apparent absence of the free dom of will. He gave the example of the donkey who, being situated between two equally sized haystacks each equally distant, should by all means die when the donkey cannot choose which haystack to go and eat. Since then a popular expression, Buridan’s donkey, has arisen. A person, irresolute in a choice or vacillating between two equivalent situations, is called “Buridan’s donkey.” 2.16.2. A Psychological Experiment We are quite confident that repeatedly throwing a coin inevitably results in 50% heads and 50% tails. American psychologist Vladimir Lefevr began to re flect upon whether one can apply this result to psychology. For this reason, he carried out the following experiment. He asked a person to divide a pile of string beans into two piles; the good string beans are in one of them and the bad string beans in the other. String beans in the initial pile are all very similar to one Chapter 2 Fibonacci and Lucas Numbers 121 another, and it is difficult to give any objective criterion for their division into the “good” and “bad” string beans. It is clear that this problem is similar to throw ing a coin and we can expect the result: 50% by 50%. However, the actual result overturned the expectations: the number of the good string beans steadily built up to 62% (0.62) from the initial number of string beans. What a surpise! As we know, 62% is the golden mean of 100%! The result of the psychological experi ment proved very surprising: a person divides the string beans into two piles in the golden ratio! Vladimir Lefevre offered an original explanatory theory for this experiment. However, to understand Lefevre’s theory, it is necessary to know some basic concepts of mathematical logic and psychology. 2.16.3. What is an Implication? The name of the English mathematician George Boole (18161864) is wide ly known in modern science. In 1854 he published An Investigation of the Laws of Thought, on Which are Founded the Mathematical Theories of Logic and Prob abilities. His work produced the beginning of a new algebra, the algebra of logic, or Boolean Algebra. Boole showed for the first time, that there is an analogy between algebraic and logical operations, if we assume that logical variables take only two values — true or false (symbolized as 1 or 0). He invented a sys tem of designations and rules, which allows one to code any statements and then to operate on them as traditional numbers. Boolean algebra has three basic operations — AND, OR, NOT, which allow one to perform logical operations of conjunction, disjunction and negation on statements. In his Laws of Thought (1854) Boole finally formulated the basis of mathematical logic. However, as often happens with mathematicians, Boole wrote that he did not see the practi cal applications of his algebra. Fortunately, regarding practical applications of his algebra the great mathematician was mistaken. Modern science, in particu lar computer science, is impossible without Boolean logic, which is widely used in the analysis and synthesis of digital automatons, as well as in the solutions of logic problems on computers. Propositions can be 1 (true) or 0 (false) and Boolean Functions connecting them are the basis of Boolean algebra. A proposition is any statement that can be evaluated from the point of view of its truth or falsity. As an example, con sider the two propositions: A = “George Boole is the creator of Boolean algebra” B = “2×2=5” The first proposition is true, that is, A=1, the second proposition is false, that is, B=0. Alexey Stakhov THE MATHEMATICS OF HARMONY 122 Propositions can be Simple or Complex. We gave examples above of simple propositions. Their basic property is the fact that they can be either true or false. The complex proposition consists of several simple propositions, thus, the complex proposition is also either true or false (1 or 0). However, the truth value of the complex proposition depends on the truth values of the simple propositions that make up the complex proposition. Thus, the complex prop osition Y is a logic function (Boolean function) of the simple propositions contained within it (for example, A and B), that is, Y=ƒ(A,B). The following logic functions — Negation, Conjunction (the logic function AND), Disjunction (the logic function OR), Addition by the Module 2, etc. are the most widespread elementary Boolean functions. Boolean functions are employed in computers. Boolean functions are also used in other fields of knowledge including implication in psychology. Implication is a logical oper ation that builds up the complex proposition corresponding to the logical con nective “if … then.” Implication consists of two simple propositions, the Ante cedent which follows “if” and the Consequent which follows “then.” If A and B are the antecedent and the consequent, respectively, then the implication is the Boolean function Y=A→B that is false (Y=0) only in the case where the antecedent is true (A=1) and the consequent is simultaneously false (B=0); for the rest of the cases the implication is true (Y=1). The implication is represented by Table 2.14. Table 2.14. A truth table for implication What would be a physical applica tion of the implication? How can the A 0 0 1 1 B 0 1 0 1 implication connect two simple propo Y 1 1 0 1 sitions? We will show this in the exam ple of the statements: A = “the given quadrangle is a square” and B = “around a given quadrangle we can describe a circle.” Let us examine the complex proposition A→B = “if the given quadrangle is a square, then around the given quadrangle we can describe a circle.” There are three variants, when the proposition A→B is true (1): 1. A is true (1) and B is true (1), that is, a given quadrangle is a square and it is possible to describe a circle around it. 2. A is false (0) and B is true (1), that is, a given quadrangle is not a square, however, it is possible to describe a circle around it (it is clear that the proposition B is not true for all quadrangles). 3. A is false (0) and B is false (0), that is, a given quadrangle is not a square and it is not possible to describe a circle around it. Note that only one variant for the implication is false (0), when A is true and B is false, that is, when a given quadrangle is a square, and at the same time, it is not possible to describe a circle around it. Chapter 2 Fibonacci and Lucas Numbers 123 In usual speech a connective “if ... then” describes a relationship of cause and effect between propositions. However, in terms of logical operations the meanings of the propositions are not considered. We only consider the truth or falsity of the propositions. Therefore, we should not be confused by the seeming “nonsense” of the implications formed by the propositions that can not be connected by their contents. For example: “If the President Bush is a democrat, then giraffes are in Africa”, “if a watermelon is a berry, then gasoline is sold at the gas station.” 2.16.4. What is a Reflection? In the most general sense reflection maybe though of as the human soul pondering itself. There is variety of definitions for this concept. More often, a reflection is determined as the analysis of one’s own ideas and experiences; a reflection is full of doubts and uncertainties. For the first time the concept of Reflection in a contemporary sense was used by the British philosopher John Locke (16321704). From his point of view, a reflection is a special operation by a person upon his own consciousness. As a result, this operation generates ideas about one’s own consciousness. Further down, in relation to a person words “his” and “himself” stands for both genders. A person is reflecting, when one says: “I think that I think,” and so forth. According to Lefevre’s opinion, a description of reflection signifies that a person a0 can be represented symbolically by two elements: a person and his “map of consciousness,” in which the information about how the person thinks about him self is placed. This means that the “map of consciousness” contains some image Z of a person regarding himself (Fig. 2.17). There therefore appears to be a reflective struc ture, in which the person possesses a “map of con sciousness,” where the person’s image about himself is represented. Symbolically such a structure can be a0 a Figure 2.17. A person with expressed by the formula A = a01 , where a0 is the person himself (how the “absolute observer” can see his “map of consciousness” Z a person from outside), and a1 is the image of a person about himself. On the other hand, it is possible toaconsider the reflection of the a 2 second order A = a01 , where a1 is the image of a person thinking about himself, and a2 is the image of a person about his own thoughts (Fig. 2.18). a 1 a0 Now, let us introduce one more intriguing Figure 2.18. A reflective word, Intention (from the Latin word intentio — structure of the second order Z Z Alexey Stakhov THE MATHEMATICS OF HARMONY 124 aspiration). This word means an aspiration, a purpose, a direction or an orienta tion of consciousness, wish, and feelings towards any subject. In the expression a0a1 the symbol a0 means the intention of the person and the symbol a1 means the intention of his image. We can estimate the intentions and the acts of the person as positive (1) or negative (0) as done in Boolean algebra. A negative estimation of oneself means the presence of guilt. A negative estimation of the partner means the feeling of the partner’s guilt and selfcondemnation. Let us consider the expression: a a 2 (2.77) A = a01 . a2 Here, the power A1 = a1 means the image of the person in his representa tion. The appraisal A1=0 means that a person estimates himself negatively, vice versa, A1=1 means that a person estimates himself positively. The power a2=1 is a representation of what a person thinks about how he estimates him self. The appraisal a2=1 means that a person has a positive selfappraisal. The appraisal a2=0 means that a person has a negative selfappraisal. Lefevre started to study this kind of selfreflection (2.77), that is, the reflec tion of the second order, where a1 is a representation (a thought) of a person about himself, a2 is a representation (a thought) of a person about his own thought. Lefevre’s important idea consists of the fact that the appraisal of a one’s self and negative or positive determination of this appraisal is fulfilled by a person without any efforts from the person’s consciousness, that is, each per son seems to have a “reflective computer” that automatically makes these self appraisals according to the formula (2.77). Now consider the values taken by the function (2.77) with dependence on the variables a0, a1, a2. With this purpose in mind, we first take into account the expression. We can write the numerical values of this expression with de pendence on the values of a1 and a2 as follows: 11=1; 01=0; 10=1; 00=1. (2.78) These expressions do not raise any objections from the mathematical point of view. In mathematics, the expression 00=1 is accepted as “correct” by definition. Now compare the expressions (2.78) with the truth table for implication a (Table 2.13). It follows from the comparison that the expression A1 = a1 2 sets the implication A1=a2→ a1. If we consider the values of the function (2.77), we can conclude that this function creates the implication of a more complicated kind, namely: (2.79) A=(a2→ a1)→ a0. Note that the expression (2.79) is Lefevre’s key idea! Let us consider a truth table for this logical function (2.79). Chapter 2 Fibonacci and Lucas Numbers 125 Table 2.15. A truth table for Lefevre’s function ao 1 1 1 1 0 0 0 0 a1 1 0 1 0 1 0 1 0 a2 1 1 0 0 1 1 0 0 А 1 1 1 1 0 1 0 0 Analysis of Table 2.15 results in very interesting conclusion. The total num ber of possible values for Lefevre’s function (2.79) is equal to 8. Five of them take the value 1 (a positive estimation) and the remaining 3 values take the value 0 (a negative estimation). Note that 8, 5 and 3 are adjacent Fibonacci numbers! This means that according to Table 2.15, our “reflective computer” automatically (that is, without participation of our consciousness) gives a Fi bonacci ratio between positive (5) and negative (3) estimations. This 5:3 ratio is a Fibonacci approximation to the golden mean of 1.618! That is, Lefevre’s theory explains an occurrence of the golden mean in the above psychological experiment with the string beans, that is, a person automatically, without the participation of his consciousness, tends to make a choice according to the golden mean principle! This paradoxical conclusion can have very unexpected applica tions to the socalled “behavioral” sciences. The theory of Elliott’s wave that has actively developed in American science is one such theory. 2.17. Elliott Waves However, if Lefevre’s theory is true, that is, all processes of the behavior of a person and his decisions are based on (or related to) the “Golden Mean Princi ple,” then perhaps Fibonacci numbers and the golden mean can be discovered in such “behavioral systems,” such as in an economy that is a function of human behavior through the decisions made by many people. Here we may be in for a big surprise! In the first half of the 20th century the American bookkeeper and econ omist Ralph Elliott developed an original theory about stock market price fluctu ations. This research resulted in the modern science of Elliott Waves. 2.17.1. Ralph Nelson Elliott Let us start from who was Elliott and how did he come by his discovery? Ralph Nelson Elliott (18711948) was one of the brightest representatives of the “American Renaissance.” Being an accountant by education, he special Alexey Stakhov THE MATHEMATICS OF HARMONY 126 ized in reorganizing and revitalizing large companies such as exportimport houses and railroads in the U.S. and Central America. In 1924 the U.S. State Department appointed Elliott to serve as Chief Accountant for Nicaragua where the U.S. had large economic ties. Later, Elliott wrote a foreign policy proposal based upon his longterm experience in Central and South America and submitted it to the State De partment. The essential ideas of his proposal were reflected in the program “Good Neighbor” as elaborated by Roosevelt’s administration, and used later in the more recent programs of the World Bank. In the 1930s Elliott turned his attention toward studying the stock mar ket. In 1938 at the age of 67, he published his first monograph about market models, The Wave Principle. In the next year, he published a series of articles in Financial World magazine that detailed his discovery. In 1946 he completed a large book, Nature’s Law, in which he expanded his 1938 essay on the connec tion between the Wave Principle, the golden mean and the stock market. 2.17.2. Rhythm in Nature The proposition that the Universe is subordinated to some general Laws returns us to Pythagoras and Plato, and the origin of this great universal recog nition in modern science. It is obvious that without the Laws of Nature we would have chaos. A strict ordering and constancy of Nature follows. This is confirmed by a surprising periodicity and repetition in all of Nature’s processes. A person is a natural object similar to the Sun or the Moon. One’s activity can be expressed by the language of numbers and is subject to scientific analy sis. Human activity, for example the heart’s activity, could be considered from the point of view of rhythmic processes. A person’s heart beats uniformly (about 60 beats per minute). The heart acts as a cylinder piston by drawing in and then pushing out the blood. Blood pressure changes during this cardiac cycle. It reaches its greatest value in the left ventricle at the moment of compression (systole). In arteries (during systole) the blood pressure is reaching its maxi mum value, equal to 115125 mm Hg. At the moment of the cardiac reduction (diastole), the pressure is decreasing until there is 7080 mm Hg. The ratio of the maximum (systolic) pressure to the minimum (diastolic) pressure is equal, on the average, to 1.6, which is a close approximation to the golden mean. Is this a mere random coincidence, or does it reflect some objective regularity of the organized cardiac activity? The heart beats continuously from a person’s birth up to his death. And its activity should be optimal and be subordinate to the selforganizational laws of biological systems. Because the golden mean is one of Chapter 2 Fibonacci and Lucas Numbers 127 the criteria of selforganizing systems, one might naturally expect the cardiac cy cle to be subordinate to the golden mean law. This hypothesis underlies the mam mal cardiac research activity of Russian biologist Tscvetkov [39]. Judgments about heart activity are done through an electrocardiogram, displaying a curve that reflects cardiac performance (Fig. 2.19). In a cardiogram, a comparison is made of two time intervals of different duration corresponding to the heart’s systolic (t1) and diastolic (t2) activity. Tscvetkov found that there exists the optimal (golden) frequency for people and other mammals; here, the durations of systole, diastole and the full cardiac cycle (T) are in golden mean proportion, that is, T: t2= t2: t1. For example, for a person this golden frequency is equal to F ORS T F ORS T 63 heart beats per minute, and 94 beats Figure 2.19. Human electrocardiogram per minute for dogs. Tscvetkov found that if we take the middle blood pressure in the aorta as the measurement unit, then the systolic blood pressure is 0.382, and the dias tolic pressure is 0.618, that is, their ratio corresponds to the golden ratio (0.618:0.382=1.618). It means that cardiac performance in the timing cycles and blood pressure variations are optimized according to the same principle, the law of the golden mean. If we move on from the investigation of the person as a biological unit to the consideration of our social and economic activity, we find that human ac tivity follows certain laws, which force the social and economic processes to repeat in the form of a certain set of waves or impulses. The best way this idea can be shown is through the example of stock exchange processes. 2.17.3. Elliott’s Wave Principle Let us now return to a consideration of the main ideas concerning the stock market that underlie the Elliott wave theory: 1. Natural Laws embrace the most important element of all, namely, tim ing. They are not themselves some simple system or method for playing the market. However, as expressed in market phenomena they appear to mark the expression of all human activities. The application of these laws to forecasting may have a revolutionary character. Alexey Stakhov THE MATHEMATICS OF HARMONY 128 2. This law can only be discerned when the market is viewed and ana lyzed as a product of human decisions. Simply put, the stock market is a human creation and therefore reflects human idiosyncrasy. 3. All human activities have three distinctive features — pattern, time and ratio — each of which appears to be based upon the Fibonacci series. According to the Elliott Wave Theory, stock prices can be simulated by using a specific wave that is connected with Fibonacci numbers. Elliott’s main wave is shown in Fig. 2.20. More specifically, Elliott believed that the main wave is simu lated by distinct upward and downward a 5 movements. For example, the ascending b 4 part of the wave consists of 5 movements, c three movements up and 2 movements 3 down (3 and 2 are of course adjacent Fi 2 bonacci numbers); the descending part of 1 the wave consists of 3 movements, 2 move ments down and 1 movement up (again ad Figure 2.20. Elliott Waves jacent Fibonacci numbers). The movements 1, 3 and 5 represent the Impulses in a Major Bull Move ment. The movements 2 and 4 represent the Corrective Movements in a major bull movement. The movements A, B and C define a Minor Bear Wave; here, the movements A and C represent the descending movements of the minor bear wave, while B represents one ascending movement of the minor bear wave. The major waves determine the major trends of the market, and the minor waves determine the minor trends. Elliott carefully examined the market’s waves, and found that the golden mean and Fibonacci numbers play an im portant role in the pressures and trends of the stock market. Elliott proposed that the main wave exists at many levels; this means that new subwaves could appear within the primary wave. To clarify, this means that the chart above (Fig. 2.20) represents only the primary wave pattern. How ever, the same kind of wave occurs, for example, between the points 2 and 4. The diagram below (Fig. 2.21) shows how primary waves could be broken down into smaller waves. We can see in Fig. 5 2 4 2.21 that the smaller waves are subor a Bc 1 3 35 dinate to one and the same principle, 5 b b A 3 the golden mean principle. C a c 1 2 4 4 Trading which is based upon the 2 Elliott Wave patterns is quite simple. 2 A trader identifies the main wave as a supercycle and then acts accord Figure 2.21. The smaller Elliott Waves Chapter 2 Fibonacci and Lucas Numbers 129 ing to the Elliott Wave model, that is, increasing or decreasing the sales be cause market behavior is subordinated to the golden mean principle. This goes on in progressively shorter cycles until the cycle is complete. Of course, real market processes are more complex than the Elliott model. Even if we do not agree with some of Elliott’s findings, the idea is of great methodological interest. Many scientists, especially in the U.S., are convinced Elliott supporters. The American scientist Robert Prechter is the most fa mous follower of Elliott’s ideas. In 1999, he published a book [143], which is dedicated to the development of the Elliott Wave Principle, and organized Elliott Wave International to promote Elliott’s ideas. Prechter made the following very ambitious statement: “(R.N. Elliott’s) Wave Principle is to sociology what Newton’s laws were to physics” [143]. Time will tell whether Prechter’s comparison of the Elliot Wave Princi ple to Newton’s Laws has merit. However, one thing is doubtless: thanks to the activities of Elliott and his followers, a theory of modern sociology and market economics was added with deep scientific significance. According to this concept, Fibonacci numbers and the golden mean are not only behind the growth of the pinecone or seashell, but also determine the laws of human behavior, and through them the laws of the stock market! 2.18. The Outstanding Fibonacci Mathematicians of the 20th Century The 20th century is characterized by an increasing interest in Fibonacci numbers and the golden section. Two scientific works, the book Aesthetics of Proportions in Nature and Art [11] by the French scientist Matila Ghyka and the book Proportionality in Architecture [10] by Russian architect Professor G.D. Grimm were very important achievements in the “golden” area in the first half of the 20th century. In addition, considerable results in this area were obtained in the 20th Cen tury by Danish mathematician Willem Abraham Wythoff and Belgian ama teur mathematician Edouard Zeckendorf. 2.18.1. Willem Abraham Wythoff Many specialists in combinatorial analysis and number theory know about the Wythoff game. However, few know anything about the man for whom this game was named. W. A. Wythoff was born in Amsterdam in 1865, the son of a Alexey Stakhov THE MATHEMATICS OF HARMONY 130 sugar refinery operator, and obtained his Ph.D. in mathematics in 1898 from the University of Amsterdam. Wythoff’s game was described in his article A Modifi cation of the Game of Nim [Nieuw Archief voor wiskunde, 2 (190507), 199202]. Dr. Wythoff described his famous game as follows: “The game is played by two persons. Two piles of counters are placed on the table, the number of each pile being arbitrary. The players play alternately and either take from one of the piles an arbitrary number of counters or from both piles an equal number. The player who takes up the last counter, or counters, wins.” The solutions to Wythoff’s game that are based on Fibonacci numbers can be found in numerous articles in The Fibonacci Quarterly, and also in the fa mous article The Golden Section, Phyllotaxis, and Wythoff’s Game by Coxeter (Scripta Mathematica, 19 (1953) 135143). 2.18.2. Edouard Zeckendorf Many number theorists know about Zeckendorf sums, however, here again, few know anything about the man for whom these sums are named. Edouard Zeckendorf was born in Liиge, Denmark. In 1925, he qualified as a medi cal doctor at the University of Liиge and then became a Belgian army officer. He also obtained a license for dental surgery some time prior to 1930. Mathematics nevertheless was Zeckendorf’s main pas Edouard Zecken dorf (19011983) sion. In 1939, he published the article devoted to Zecken dorf sums. According to his theorem, each positive integer has a unique repre sentation as a sum of Fibonacci numbers, though two adjacent Fibonacci num bers are actually never employed. Research on Zeckendorf sums became the subject of many articles published in The Fibonacci Quarterly. 2.18.3. Verner Emil Hoggatt The mathematical organization of the Fibonacci Associ ation was founded in 1963 by a group of American mathe maticians making it one of the most outstanding events in the history of Fibonacci number theory. Beginning of 1963, the Fibonacci Association began to issue The Fibonacci Quar terly. The American mathematicians Verner Emil Hoggatt (19211981) and Alfred Brousseau (19071988) were Verner Emil Hoggatt founders of the Fibonacci Association. Verner Hoggatt, along (19211981) Chapter 2 Fibonacci and Lucas Numbers 131 with Brother Alfred Brousseau, published the first volume of The Fibonacci Quar terly in 1963, thereby founding the Fibonacci Association. The April 4, 1969 issue of TIME magazine reported the phenomenal growth of the Fibonacci Association. In the same year, Houghton Mifflin published Verner Hoggatt’s book, Fibonacci and Lucas Numbers [16], perhaps the world’s best introduction to Fibonacci num ber theory. Howard Eves wrote that “during his long and outstanding tenure at San Jose State University, Vern directed an enormous number of master’s theses, and put out an amazing number of attractive papers. He became the authority on Fibonacci and related numbers.” 2.18.4. Alfred Brousseau (19071988) Alfred Brousseau was another outstanding person involved in the orga nization of the Fibonacci Association. Brother Alfred belonged to the reli gious order Fratres Scholarum Christianarum that translates as Brothers of the Christian Schools, or simply, The Christian Brothers. Brother Alfred was an avid photographer. He made a col lection of some 20,000 slides of California wildflowers. Im ages of more than half of these and other collections by Broth er Alfred are preserved in the form of widely used websites at the University of California, Berkeley. In 1969, Time magazine featured two founders of the Fibonacci Association in an article titled The Fibonacci Numbers. Brother Alfred was pictured in the Time article holding a pineapple, one of the best known representatives of phyllotaxis. The article re Alfred Brousseau ferred to the many natural applications of Fibonacci numbers: for (19071988) example, male bees reproduce “fibonaccically,” and Fibonacci numbers occur in the formations visible in many sunflowers, pine cones, and leafpositions on branches of trees. Brother Alfred recommended that people who learn about Fibonacci num bers should focus their attention on the aesthetic pleasure involved in it. 2.18.5. Steven Vaida In 1989, twenty years after Verner Hoggat’s book, the publishing house of Ellis Horwood Limited published the book Fibonacci & Lucas Numbers and the Golden Section [28] by Prof. S. Vaida. The book attracted wide attention from Fibonacci mathematicians, as it is considered to be one of the best mathemat ical books on the subject. Who is the author of this book? From his brief sci entific biography, we learn that the mathematician Steven Vaida was Profes sor of Mathematics of University Sussex (England) at the time of the book’s Alexey Stakhov THE MATHEMATICS OF HARMONY 132 publication. He obtained a Doctorate in Philosophy at Vienna University. In 1965, he began to work as Professor of operational researches at Birmingham University. It is also noteworthy that he was an honorary member of many mathematical organizations, in particular, the Mathematics and Statistics In stitute of the Society of Operational Researches (England). 2.18.6. Herta Taussig Freitag Among the modern Fibonacci mathematicians, it is necessary to name Pro fessor Herta Taussig Freitag, another member of the Fibonacci Association. She was born on December 6th, 1908 in Vienna, Austria and died January 25th, 2000 in Roanoke, Virginia. Herta Freitag graduated with a Ph.D. from Columbia Uni versity in 1953. Her colleagues at the Fibonacci Association named Herta Freitag the Queen of the Fibonacci Association. Over the years, she attended and presented a paper at every In ternational Conference of Fibonacci numbers starting with the first Conference in 1984. It is curious to note that Professor Fre itag delivered the lecture Elements of Zeckendorf Arithmetic (co Herta Taussig author G.M. Phillips) at the 7th International Conference on Freitag (19082000) Fibonacci Numbers and Their Applications. The appearance of this lecture title is rather distinctive. It testifies to the fact that Fibonacci mathe maticians came close to the creation of the new computer arithmetic based on Fibonacci numbers and Zeckendorf sums. This became a central subject of the Soviet scientific and engineering developments in the 1970s and 1980s. 2.19. Slavic “Golden” Group The outstanding Hungarian geometer Janos Bolyai, one of the creators of NonEuclidean geometry, once wrote the following remarkable words: “For ideas, as well as for plants, the time comes, when they mature in their various locales, just as in the spring the violets blossom wherever the Sun shines.” The history of science should tell us why the 80s and 90s became the his torical period in which interest in Fibonacci numbers and the Golden Section began to peak. Particularly in this period, scientists of different disciplines put forward hypotheses concerning the applications of the golden mean and made discoveries that have fundamental significance for the development of science and its many branches. Chapter 2 Fibonacci and Lucas Numbers 133 The small brochure Fibonacci Numbers [13] published in 1961 by the Rus sian mathematician Nikolay Vorobyov resulted in an expanding mathemati cal interest in Fibonacci numbers. The brochure was reissued repeatedly and translated into many languages. First of all, it is necessary to mention the primary direction that arose in the 1970s in Soviet science. It is well known that Fibonacci’s problem of “rabbit reproduction” became a source for research into Fibonacci numbers. However, the “rabbit reproduc tion problem” is not the only problem suggested by Fibonacci. He also proposed the wellknown problem of Choosing the Best System of Standard Weights. This problem is named the Weighing Problem or BashetMendeleev Problem in Russian historicmath ematical literature. In 1977, the author of this book published the book Introduction into Algorithmic Measurement Theory [20] and in 1979 the brochure Algorithmic Mea surement Theory [21]. These works generalize the BashetMendeleev problem and provide a new measurement theory based on Fibonacci numbers, namely Algorithmic Measurement Theory. Published at the end of the 70s they helped determine the ap plied nature of research in the Fibonacci field (Fibonacci codes, Fibonacci arithmetic, Fibonacci computers) that began to develop in Soviet science. The practical direction of Fibonacci research by the Soviet scientific school is distinctively different from the direction of the American Fibonacci Association. In 1984, two books were published dedicated to the golden section. The Byelorussian philosopher Eduard Soroko in his book Structural Harmony of Systems [25] made a courageous attempt to revive for modern science the Pythagorean idea of the numeric harmony of the Universe, proposing the so called Law of Structural Harmony of Systems that is expressed with the help of the generalized golden sections, namely, the golden рsections [20]. In the same year Stakhov’s book Codes of the Golden Proportion [24] was published. This book developed number systems with irrational bases or codes of the golden proportion that are a generalization of Bergman’s number system [86]. One year later in 1985, the book Aesthetic Foundations of Ancient Egyp tian Art [145] by Russian art critic Natalia Pomerantseva was published; the book demonstrates quite convincingly the role of the golden section in An cient Egyptian culture. The year 1986 was marked by the publication of two books on the Golden Section. In Poland the EnergyGeometric Code of Nature [26] by Polish scien tist and journalist Jan Grzedzielski was published. This book, probably for the first time, uncovered the physical sense of the golden mean as the main code of the Universe, as the proportion of thermodynamic equilibrium in selforganiz ing systems. In the same year a teacher at the Kiev State Art Institute, Kovalev, published a manual for artists called The Golden Section in Painting [29]. Alexey Stakhov THE MATHEMATICS OF HARMONY 134 The beginning of the 1990s was characterized by the publication of two more books on the Golden Section. The first was The Golden Section: Three Approaches to the Nature of Harmony (1990) [146]. The book was authored by three representatives of Russian art, the architect Shevelev, the composer Marutaev and the architect Shmelev. The abstract states: “This book is dedi cated to the theoretical substantiation of the phenomenon of the golden sec tion, one of the universal laws of Nature.” The popular book The Golden Proportion [31] by Nikolay Vasyutinsky, a Ukrainian researcher, chemist, geologist, metallurgist and mechanic, was of considerable importance of 1990. Written with great proficiency, it states: “the manifestation of the Golden Proportion laws in architecture, music, poetry, chemistry, biology, botany, geology, astronomy, and engineering sciences are herein described.” At the beginning of the 1990s, it became clear that Slavic science (Ukraine, Russia, Belarus, and Poland) had developed an immensely powerful group of re searchers formed of representatives of various sciences and arts, and authors of very original books on the golden section. The idea arose to unite these research ers together and create a Slavic golden scientific society. The First International Seminar of The Golden Proportion and Problems of System Harmony was held in Kiev in 1992 under the scientific supervision of Professor Stakhov, the author of this book. The Belarussian philosopher Eduard Soroko (Minsk), the Ukrainian architect and art critic Оleg Bodnar (Lvov), the Ukrainian mathematician and economist Ivan Tkachenko, the Russian mechanical engineer Victor Korobko (Stavropol), the Ukrainian physician and anatomist Pavel Shaparenko (Vinnit sa), the Ukrainian chemist Nikolay Vasjutinsky (Zaporozhye), and the Polish journalist Jan Grzedzielski (Warsaw) participated as members of the group’s or ganizing committee. This group of highly respected scientists became the skele ton of the Slavic scientists’ association called the Slavic Golden Group. The Second International Seminar of The Golden Proportion and Problems of System Harmony was held in 1993 in Kiev. Then on the initiative of Profes sor Korobko, the Seminar continued its work in 1994, 1995 and 1996 in Stavropol, Russia. The Seminars stimulated research in the golden section in a variety of fields. During the 1990s the Slavic Golden Group published a number of very interesting books on the subject. In 1994, the book The Golden Section and NonEuclidean Geometry in Na ture and Art [37] was published by Oleg Bodnar. A new geometrical theory of phyllotaxis (the Law of Spiral Biosymmetry Transformation) discovered by Bodnar became the primary outcome of the book. Chapter 2 Fibonacci and Lucas Numbers 135 The end of the 20th century was very successful for the Slavic Golden Group. In 1997 Heart, Golden Section and Symmetry [39] by Russian biologist Tsvetk ov was published. The book is a summary of the author’s longterm research in the area of Golden Section applications to cardiac activity in mammals. In this book the set of Golden Sections in the different structures of the cardiac cycle is found. Tsvetkov also demonstrated the role of the Golden Section and Fibonac ci numbers in optimization of the heart activity (minimization of energy con sumption, in the blood, muscle and vascular matter) of mammals. In 1998, the book The Golden Proportion and Problems of System Harmony [43] was published by Professor Korobko (Russia), an active member of the Slavic Golden Group. The book contains considerable information concerning the application of the golden section throughout Nature, Science and Art. The great strength of the book is that it can serve as a manual for university and college teachers of postgraduate students in both the engineering sciences and the liberal arts. In fact, the Russian Association for Building Universities rec ommended it as a manual for all students in colleges and universities. In 1999, Alexey Stakhov with assistance of Vinancio Massingue and Anna Sluchenkova published Introduction into Fibonacci Coding and Cryptography [44]. This book presented a new coding theory based on Fibonacci matrices. In 2000, the famous Russian architect Shevelev [46] published the book The Metalanguage of Living Nature. The beginning of the 21st century has been characterized by increasing scientific activity by the Slavic Golden Group. In 2001, Alexey Stakhov and Anna Sluchenkova created a website Museum of Harmony and the Golden Section www.goldenmuseum.com. The uniqueness of this site in its bilingual ism (Russian/English) and big concentration of exclusive information about Golden Section and its applications. In 2003, the Slavic Golden Group organized the international conference Prob lems of Harmony, Symmetry and the Golden Section in Nature, Science and Art in the Ukrainian city of Vinnitsa. Here the great contribution of the Slavic scientists to the development of the theory of the golden section and its applications was recog nized in the 38 books written on the golden section. According to the Conference resolution, the Slavic Golden Group was transformed into the International Club of the Golden Section. This group stimulated Slavic researchers to further develop their creative work. Included amongst the many interesting books written by this presti gious group of leading Slavic scientists, were the following very popular ones: 1. Bodnar, O.J. The Golden Section and NonEuclidean Ggeometry in Sci ence and Art (2005) (in Ukrainian) [52]. This book is the second edition of Bodnar’s preceding book [37]. Alexey Stakhov THE MATHEMATICS OF HARMONY 136 2. Petrunenko, V.V. The Golden Section in Quantum States and its Astro nomical and Physical Manifestations (2005) (in Russian) [53]. 3. Soroko, E. M. The Golden Section, Processes of Selforganization and Evolution of System. Introduction into General Theory of System Harmony (2006) (in Russian) [[56]. This book is the second edition of Soroko’s preceding book [25]. 4. Stakhov, A.P., Sluchenkova, A.A., and Scherbakov, I.G. The da Vinci Code and Fibonacci Series (2006) (in Russian) [55]. Together with the American Fibonacci Association, the Slavic Golden Group greatly influenced the development of contemporary research in the field of the golden section and Fibonacci numbers, and their applications. 2.20. Conclusion The Fibonacci and Lucas numbers are two remarkable numerical sequences, which are becoming widely known in modern science. Fibonacci numbers were introduced into mathematics in the 13th century by the famous Italian mathema tician Leonardo of Pisa (Fibonacci) as the solution to the Fibonacci rabbit repro duction problem. Fibonacci and Lucas numbers have many interesting mathe matical properties. If we take the ratio of two adjacent Fibonacci numbers Fn/Fn1 and direct n towards infinity, the ratio approaches the golden mean in the limit. The discovery of this mathematical property is attributed to Johannes Ke pler. In the 17th century Kepler’s contemporary, the great astronomer Giovanni Cassini, proved the remarkable Cassini formula connecting three neighboring Fibonacci numbers. If we calculate the numerological values of all the terms of the Fibonacci series, then we will find that there is an intriguing periodicity, equal to 24, in this series of numerological values. The French 19th century mathemati cians Lucas and Binet made great contributions to the development of Fibonacci number theory. Lucas introduced Lucas numbers into mathematics and Binet derived the famous mathematical formulas (Binet formulas) which connect Fi bonacci and Lucas numbers with the golden mean. The Fibonacci and Lucas num bers show themselves throughout Nature’s structures. The botanic phenomenon of phyllotaxis is the best recognized amongst them. However, during the last few decades, Fibonacci numbers have been revealed in the genetic code (Fibonacci Resonances), in psychology (Lefevre’s experiments), and the extension of market processes (Elliott Waves), and numerous other areas and disciplines. Chapter 3 Regular Polyhedrons 137 Chapter 3 Regular Polyhedrons 3.1. Platonic Solids 3.1.1. Regular Polygons People show an interest in regular polygons and polyhedrons (i.e. polyhedra) throughout life – from the twoyearold child, playing with wooden cubes, to the mature mathematician. Some regular and semiregular solids appear in Nature in the form of crystals, others as viruses seen only by using an electron microscope. What is a polygon? Recall that geometry may be defined as a science of space and spatial figures – twodimensional and threedimensional. A two dimensional geometric figure can be defined as a set of line segments limiting some part of a plane. Such a planar figure is called a Polygon. Scientists have been interested for some time in the Ideal or Regular poly gons, that is, the polygons having equal sides and equal angles. The idea of “regularity” and “ideal geometric figures” is very old in geometry. It dates back to the ancient Greek mathematicians and philosophers. The equilateral triangle is considered to be the simplest regular polygon be cause it has the least number of sides necessary to limit part of a plane. The square (four sides), pentagon (five sides), hexagon (six sides), octagon (eight sides), deca gon (ten sides) and so on to gether with the equilateral tri angle (Fig. 3.1) provide a gen eral picture of the regular poly Equalateral Cube Pentagon Hexagon gons. It is obvious that there are no theoretical restrictions on the number of sides of a regular Octagon Nonagon Decagon polygon, meaning there are an Heptagon Figure 3.1. Regular convex polygons infinite number of polygons. Alexey Stakhov THE MATHEMATICS OF HARMONY 138 Note that all regular polygons in Fig. 3.1 are Convex. What is a convex geo metric figure? The Convex geometric figure is the opposite of a Concave geo metric figure. To understand the distinction between convex and concave geometric figures, we examine two famous figures, the pentagon (Fig. 3.2a) and pentagram (Fig. 3.2b). The regular pentagon in Fig. 3.2a is a planar convex figure, while the pentagram in Fig. 3.2b is a planar concave figure. The notion of convex and concave geometric a) b) figures should be intuitively clear. Below we clar Figure 3.2. Regular pentagon (a) ify these geometric concepts in greater detail when is a convex figure, pentagram (b) is a concave figure we begin the study of regular polyhedra. The first textbook of geometry, Euclid’s Elements, assumed convexity with out providing a precise definition. Convex polygons have an interesting prop erty connected with the interior angles of polygons. We know that the three angles in any triangle always total 180°. As the equilateral triangles have equal angles, each of its angles must be equal to 60° (60+60+60=180). The angles of any quadrangular polygon add to 360°. As the angles of a regular polygon are equal, each angle of the regular quadrangular polygon (square) is equal to 90° (90+90+90+90=360). The angles of any fivesided polygon add to 540°. It is clear that each angle of the regular pentagonal polygon (pentagon) is equal to 108° (108+108+108+108+108=540). How can we calculate the interior angle of any regular ngon? Take any vertex of a regular ngon and draw from this vertex all possible diagonals within the poly gon. This process divides the regular ngon into n–2 (n minus 2) nonintersecting triangles. As three angles of a triangle total 180°, then the interior angles of any regular ngon total [180° (n2 )], a characteristic property of the convex polygon. Table 3.1 lists the numerical values for the interior angles of regular ngons. Table 3.1. Interior angle measures in regular polygons Name Triangle Square Pentagon Hexagon Octagon Nonagon Decagon Dodecagon ... ngon Number of sides 3 4 5 6 8 9 10 12 ... n Sum of interior angles 180 360 540 720 1080 1260 1440 1800 ... 180(n2) Interior angle 60 90 108 120 135 140 144 150 ... 180(n2) /n Chapter 3 Regular Polyhedrons 139 Note that regular polygons have many lines of symmetry. This character istic is very important in their ability to build up different tessellations. We can see from Fig. 3.3 that the regular oc tagon has eight axes of symmetry: one be tween each pair of opposite vertices and one between each pair of opposite sides. The regular pentagon has five axes of symmetry: one between each vertex and its opposite side. A regular polygon with Figure 3.3. Symmetry lines of the regular octagon and the regular pentagon n sides has n axes of symmetry. 3.1.2. Regular Polyhedra A polyhedron is a “solid” threedimensional figure similar to twodimen sional polygons discussed above. Polyhedra have vertices, edges and faces. If a polyhedron has faces that are regular polygons and if at each vertex exactly the same number of faces meets, such a polyhedron is called a Regular Polyhedron. How many regular polyhedra exist? At first sight, the answer to this question is very simple: as many as there are regular polygons that are faces of regular poly hedra. However, this is not the case. To find the correct answer to this question we must divide all regular polyhedra into convex and concave types. Let us start with the convex polyhedra. Euclid’s Elements give a strict proof of the fact that only five convex regular polyhedra exist, and their faces must be one of only three types of the regular polygons: triangles, squares and pentagons (Fig. 3.4). a) b) c) d) e) Figure 3.4. Platonic Solids: (a) tetrahedron (Fire), (b) octahedron (Air), (c) hexahedron or cube (Earth), (d) icosahedron (Water), (e) dodecahedron (Universal Mind) Many books are devoted to the theory of polyhedra. Polyhedron Models by English mathematician M. Wenninger is the best known amongst them. The Russian translation of this book was published in 1974 [147]. Bertrand Russell’s statement was chosen as an epigraph to this book: “Mathematics possesses not only truth, but also supreme beauty … sublimely pure, and capa bly of a stern perfection such as only the greatest art can show.” Alexey Stakhov THE MATHEMATICS OF HARMONY 140 The book [147] begins with a description of the convex regular polyhedra, that is, the polyhedra that are regular polygons of one type. These polyhedra are named the Platonic Solids in honor of the ancient Greek philosopher Plato, who used regular polyhedra in his cosmology. We begin our consideration with the regular polyhedra possessing equilateral triangles as faces. The Tetrahedron is the first and simplest of them (Fig. 3.4a). The key observation is that the sum of the interior angles of the polygons that are meeting at every vertex is always less than 360 degrees. In the tet rahedron, three equilateral triangles (the sum of their interior angles is equal to 180°=3×60°) meet at each vertex; thus, their bases create a new equilateral triangle. The tetrahedron has the least number of faces among the Platonic Solids and there fore it is the threedimensional analog to the equilateral triangle (which of course has the least number of sides among the regular polygons). The Octahedron (Fig. 3.4b) is the next spatial geometric figure based on equilateral triangles. In the octahedron, four equilateral triangles (the sum of their interior angles is equal to 240°=4×60°) come together at one vertex; as a result, a pyramid with a quadrangular base arises. If one connects two such pyramids by their bases, then the symmetric figure with eight triangular fac es, called the Octahedron, appears. Now, we can try to connect 5 equilateral triangles at one vertex (the sum of the interior angles is equal to 300°=5×60°). As a result, we obtain a spatial geometric figure with 20 triangular faces called the Icosahedron (Fig. 3.4d). A square is the next regular polygon (with interior angle of 90°). If we unite 3 squares at one vertex (the sum of their interior angles equaling 270°=3×90°) and then add to this figure three new squares, we obtain a per fect geometric figure with 6 faces called a Hexahedron or Cube (Fig. 3.4c). Finally, there is one more regular polyhedron to construct based on the use of a pentagon with the interior angle 108°. If we collect 12 pentagons so that 3 regular pentagons come together at each vertex (the sum of their inte rior angles is equal to 324°=3×108°), then we obtain the next Platonic Solid with 12 pentagonal faces called the Dodecahedron (Fig. 3.4e). The hexagon is the next regular polygon after the pentagon. It has an interior angle 120°. If we connect 3 hexagons at one vertex, we obtain a sur face because the sum of their interior angles is equal to 360°=3×120°. This means that it is impossible to construct a threedimensional geometric fig ure from hexagons alone. Other regular polygons after the hexagon have interior angles greater than 120°. This means that we cannot construct spa tial geometric figures from them. It follows from this examination that there are only 5 convex regular polyhedra, the faces of which are limited to equilateral triangles, squares and pentagons. Chapter 3 Regular Polyhedrons 141 There are surprising geometrical connections between all regular polyhe dra. For example, the cube (Fig. 3.4c) and the octahedron (Fig. 3.4b) are dual to one another, that is, they can be obtained from each other, if the centroids of the faces of the first figure are taken as the vertices of the other and conversely. Similarly, the icosahedron (Fig. 3.4d) is dual to the dodecahedron (Fig. 3.4e). The tetrahedron (Fig. 3.4a) is dual to itself. The fact of the existence of only five convex regular polyhedra is surprising because the number of regular poly gons is infinite! This fact was proven in Euclid’s Elements. 3.1.3. Numerical Characteristics of Platonic Solids Before continuing, let us collect some data about the regular polyhedra. Let • m be the number of polygons meeting at one vertex, • n be the number of vertices of each polygon, • f be the number of faces of a polyhedron, • e be the number of edges of a polyhedron, and • v be the number of vertices of a polyhedron. The values of these numbers for Table 3.2. Numerical characteristics of regular polyhedra each of the regular polyhedra are list ed in Table 3.2. n m f e v Further, our aim is to show that for Tetrahedron 3 3 4 6 4 3 4 8 12 6 any pair of numbers n and m the values Octahedron Icosahedron 3 5 20 30 12 of the other parameters f, e, and v are Hexahedron 4 3 6 12 8 uniquely determined. As two faces come Dodecahedron 5 3 12 30 20 together at one edge, we can write: e=nf / 2. Next, as m faces come together at each vertex, we can write: (3.1) v=nf / m. (3.2) It is apparent from Table 3.2 that for all five regular polyhedra we have: f=2+ev. The result (3.3) is known as Euler’s Polyhedron Theorem. (3.3) 3.1.4. The Golden Section in the Dodecahedron and Icosahedron The dodecahedron and its dual the icosahedron (Fig. 3.4d, e) take up a special place among the Platonic Solids. First of all, it is necessary to empha size that the geometry of the dodecahedron and icosahedron is directly con nected with the golden section. The faces of the dodecahedron (Fig. 3.4e) are Alexey Stakhov THE MATHEMATICS OF HARMONY 142 pentagons based on the golden section. If we look carefully at the icosahedron (Fig. 3.4d), we can see that five triangles come together at each vertex of the icosahedron; here, their external sides produce a pentagon based on the gold en section. Already these facts demonstrate that the golden section plays an essential role in these two Platonic Solids. A duality of the dodecahedron and the icosahedron is expressed in the fact that the number of the dodecahedron faces (f=12) is equal to the number of the icosahedron vertices (v=12) and the number of the icosahedron faces (f=20) is equal to the number of the dodecahedron vertices (v=20), but they both have the same number of edges (e=30). There is another numerical characteristic that unites the dodecahedron and the icosahedron: the number of planar angles on the surface of both spatial figures is equal to 60. As the number of vertices (and sides) n of the regular polygons (triangle and pentagon) that are the faces of the icosahedron and dodecahedron are equal to 3 and 5, respectively, and the number of faces of the icosahedron and dodecahedron are equal to 20 and 12, respectively, it follows that the following formula for the number of the planar angles on the surface of the icosahedron and the dodecahedron are valid: 60=3×20=5×12. (3.4) There are also deeper confirmations of the fundamental role of the gold en section in the icosahedron and the dodecahedron. It is recognized that these Platonic Solids have three unique spheres. The first sphere (the in scribed sphere or insphere) is a sphere that is inserted into the Platonic Sol id and touches the centroids of its faces. Define the radius of this insphere by Ri. The second or middle sphere (midsphere or intersphere) touches the centroids of its edges. Define the radius of the midsphere by Rm. The third (external) sphere or circumsphere is circumscribed around the Platonic Solid and passes through its vertices. Define this radius by Rc. It was proven in geometry that the radius lengths of the indicated spheres for the dodecahe dron and the icosahedron with sides equal to 1 are expressed by the golden mean τ (see Table 3.3). Table 3.3. Connection of the icosahedron and dodecahedron with the golden mean τ Icosahedron Dodecahedron Rc Rm Ri W 3W W W 2 2 2 3 W 3 W 2 W 2 2 2 3W 2 2 Chapter 3 Regular Polyhedrons 143 3(3 − τ) Rc = are equal for both the icosa Ri τ hedron and the dodecahedron. Thus, if the dodecahedron and the icosahe dron have identical inserted spheres, then their described spheres are also iden tical. A proof of this mathematical theorem is given in Euclid’s Elements. Note that the ratios of the radii Table 3.4. The golden mean in the areas and volumes of the icosahedron and dodecahedron Icosahedron Outer area Volume Dodecahedron 15W 5 3 5W 6 5 3W 5W 3 2(3 W) There are other wellknown geometric relations for the dodecahedron and the icosahedron verifying their connection with the golden mean. For exam ple, if we take the icosahedron and the dodecahedron with edge length equal to 1 and compute their external areas and volumes, then they will be expressed in terms of the golden mean (see Table 3.4). Thus, there are a large number of relations obtained by the mathemati cians of antiquity, which verify the remarkable fact, that the golden mean is the main proportion of the dodecahedron and icosahedron. A connection of the golden mean with dodecahedron and icosahedron is especially interesting from the point of view of the socalled “dodecahedronicosahedron doctrine” considered below. 3.1.5. Plato’s Cosmology We mentioned above that the regular polyhedra were named Platonic Sol ids because they played a very important part in Plato’s cosmology. In Plato’s cosmology, the first four polyhedra personified four “essences” or “elements.” The Tetrahedron symbolized Fire with its top pointed upwards; the Icosahedron symbolized Water as the most “fluid” polyhedron; the Cube symbolized Earth, as the “steadiest” polyhedron; the Octahedron symbol ized Air presumably the most “aerial” polyhedron. The fifth polyhedron, the Dodecahedron, symbolized “the real World,” Plato “Universal Reason,” and was considered the main geometrical (427 347 BC) figure of the Universe or entire Cosmos. Alexey Stakhov THE MATHEMATICS OF HARMONY 144 The ancient Greeks considered harmonious relations as a basis of the Uni verse, therefore the first four “elements” were connected by the following pro portion: Earth/Water = Air/Fire. According to Plato, the atoms of the “ele ments” are in perfect consonances similar to four strings of lyre. It is pertinent to tell that such system of “elements” that includes the four elements, Earth, Water, Air and Fire, was canonized also by Aristotle. These four “elements” remained the four fundamental stones of the Universe within many centuries. Here we can use an analogy with today’s wellknown four states of substance: rigid, liquid, gaseous and plasma. Thus, the ancient Greek idea about the Harmony of the Universe was bound together with its embodiment in the Platonic Solids. Plato’s idea about the role of the regular polyhedra in the structure of the Universe influenced Euclid in his Elements. In this famous book, which over the centuries was the unique text book of geometry, a description of “ideal” lines and figures is given. A straight line is the most “ideal” line, and also the regular polygons and polyhedra are the most ideal geometric figures. It is interesting that Euclid’s Elements begins with the description of an equilateral triangle that is the simplest regular polygon and ends with a study of the five Platonic Solids. We mentioned above that the theory of the Platonic Solids was stated in the 13th, that is, final book of Eu clid’s Elements. That is why, the ancient Greek mathematician Proclus, a com mentator on Euclid, put forward the interesting hypothesis about the true pur pose for which Euclid wrote the Elements. In Proclus’ opinion, Euclid wrote his Elements in order to provide a complete theory of the construction of the “ideal” geometric figures, in particular, the five Platonic Solids. In passing he gave in the Elements some advanced achievements of the Greek mathematics necessary to give a complete theory of the “ideal” geometric figures! Thus, we can consid er Euclid’s Elements as the first historically geometric theory of the Harmony of the Universe, based on the “Golden Section” (division in extreme and mean ratio) and the Platonic Solids! 3.2. Archimedean Solids and Starshaped Regular Polyhedra 3.2.1. Archimedean Solids There are 13 semiregular convex polyhedra attributed to Archimedes. This wellknown set of the perfect geometric figures is named Archimedean or Semiregular Polyhedra. Chapter 3 Regular Polyhedrons 145 The Archimedean Solids can be divided into several groups. The first group consists of the five polyhedra that are formed from the Platonic Solids by means of the trun cation of their vertices. For the Platonic Solids the trunca tion can be made in such a manner that the new faces and the remaining parts of the old faces are regular polygons. For example, the tetrahedron (Fig. 3.4а) can be truncated so that its four triangular faces are converted into hexago Archimedes nal faces and the four new regular triangular faces are add (287 – 212 BC) ed to them. In this manner the five Archimedean Solids are obtained: Truncat ed Tetrahedron, Truncated Hexahedron (Cube), Truncated Octahedron, Trun cated Icosahedron, and Truncated Dodecahedron (Fig. 3.5). a) b) c) d) e) Figure 3.5. Archimedean solids: (а) truncated tetrahedron, (b) truncated cube, (c) truncat ed octahedron, (d) truncated dodecahedron, (e) truncated icosahedron In his Nobel lecture (1996) the American chemist Richard E. Smalley, one of the authors of the experimental discovery of fullerenes, spoke about Archimedes (287 212 BC) as the first researcher of truncated polyhedra, in particular, the Truncated Icosahedron. It is his opinion, however, that Archimedes may have as cribed to himself this discovery, even though the icosahedron was truncated long before him. All these solids were described by Archimedes, although, his original works on this topic were lost and were known only from secondhand sources. Various scientists gradually rediscovered all these polyhedra during the Renais sance, and Kepler finally reconstructed the entire set of Archimedes’ polyhedra in his 1619 book The World Harmony (“Harmonice Mundi”). How we can construct the Archi medean truncated icosahedron from the Platonic icosahedron? The answer to this question is given in Fig. 3.6. We can see from Table 3.2 that five faces converge in Figure 3.6. Construction of the each of the 12 vertices of the Platonic icosa Archimedean truncated icosahedron from the Platonic icosahedron hedron. If we truncate 12 vertices of the icosahedron by a plane, then 12 new pentagonal faces will be formed. The old triangular faces are converted into hexagonal faces. 12 new pentagonal faces to Alexey Stakhov THE MATHEMATICS OF HARMONY 146 gether with 20 hexagonal faces compose a truncated icosahedron with 32 faces. Here the edges and vertices are equal to 90 and 60, respectively. Another group of Archimedean solids consists of two quasiregular polyhe dra. A polyhedron is called Quasiregular if it consists of two sets of regular polygons, let’s say, msided and nsid ed, respectively, and it is constructed so that each polygon of the first set is com pletely surrounded by polygons of the second set. There are two quasiregular Archimedean solids named Cuboctahe dron and Icosidodecahedron (Fig. 3.7). a) b) For the cuboctahedron (Fig. 3.7a) m = Figure 3.7. Cuboctahedron (a) and 3, n = 4, and for the icosidodecahedron icosidodecahedron (b) (Fig. 3.7b) m = 3, n = 5. There are two Archimedean solids called the Rhombicuboctahedron (Fig. 3.8a) and the Rhombicosidodecahedron (Fig. 3.8b). The rhombicubocta hedron (Fig. 3.8a) consists of two kinds of faces, squares and triangles. Each square is surrounded by four squares and each tri angle is surrounded by three squares. Some believe that the rhombicosi dodecahedron (Fig. 3.8b) is the most attractive among the Archimedean sol ids. The rhombicosidodecahedron con sists of pentagons, squares and triangles. a) b) Figure 3.8. Each pentagon is surrounded by five Rhombicuboctahedron (a) squares and there is a triangle surround and rhombicosidodecahedron (b) ed by three squares. Finally, there are two socalled “snubnosed” versions of the Archimedean solids – one for the hexahedron (cube), Snub Hexahedron (Fig. 3.9a), another – for the dodecahedron, Snub Dodecahedron (Fig. 3.9b). The snub hexahedron (Fig. 3.9a) consists of six squares surround ed by triangles. Each square is surround ed by four triangles. Each triangle ad joined to a square is surrounded by two triangles. The same idea underlies a snub dodecahedron (Fig. 3.9b). It consists of 12 pentagons where each is surrounded a) b) by five squares. The gap between squares Figure 3.9. Snub hexahedron (a) and is filled by triangles. snub dodecahedron (b) Chapter 3 Regular Polyhedrons 147 3.2.2. Starshaped Regular Polyhedra In Polyhedron Models by M. Wenninger [147] we find 75 various mod els of regular and semiregular polyhedra. The Russian mathematician Ljusternak, who made much of this area, wrote: “A theory of polyhedra, in particular, the convex polyhedra, is one of the most fascinating chap ters of geometry.” The development of this theory is connected with the names of several outstanding scientists. We mentioned Kepler’s contri bution to the development of the theory of polyhedra. He wrote the etude About snowflakes where he noted the following: “Among the regular poly hedra the cube is the very first one, the beginning and the primogenitor of others, if it is permissible so to say, its spouse is the octahedron be cause it has as many vertices as the cube has faces.” Kepler published a full list of the 13 Archimedean solids and gave them the names under which they are now known. Also Kepler started to study the socalled Starshaped Polyhedra that un like Platonic and Archimedean Solids are regular concave polyhedra. In the beginning of the 19th century the French mathematician and physicist Poin sot (1777 1859), whose geometric works relate to starshaped regular poly hedra, followed Kepler’s work and discovered two new starshaped regular polyhedra. So, thanks to Kepler and Poinsot’s works, four types of the star shaped regular polyhedra (Fig. 3.10) became known. In 1812, the French math ematician Cauchy (1789 1857) proved that other starshaped regular poly hedra do not exist. Many readers may ask a question: “For what purpose is it necessary to study regular polyhedra? What benefit can we have from them?” We could answer this question by another question: “What benefit can we derive from music or poetry? Is all that is beautiful merely useful?” The models of polyhe dra presented in Figs. 3.43.10, first of all, make an aesthetic impression upon us and can be used as decorative orna ments. However, it will be shown that a wide manifestation of regular polyhedra in natural structures caused a great deal of interest in mod Figure 3.10. Starshaped regular polyhedra ern science. (KeplerPoinsot solids) Alexey Stakhov THE MATHEMATICS OF HARMONY 148 3.3. A Mystery of the Egyptian Calendar 3.3.1. What is a Calendar? A Russian proverb states: “Time is the eye of history.” Everything that exists in the Universe: the Sun, the Earth, stars, planets, known and unknown worlds – they all have spatialtemporal measurement. Time is measured by observation of the periodically repeating processes of definite duration. In remote antiquity people noted that the day is always followed by night, the seasons of the year alternating on a regular basis: after the winter the spring comes, after the spring the summer, after the summer the autumn… By searching for the cause of these phenomena, a person paid attention to the celestial heavenly bodies, the Sun, the Moon, stars, and the strict period icity of their movement in the firmament. It was the first celestial observa tions that preceded the origin of astronomy, one of the most ancient sciences. As the basis of time measurement, the astronomers had used three important astronomical phenomena: rotation of the Earth around the axis, motion of the Moon around the Earth and motion of the Earth around the Sun. The various concepts of time often depended on the physical phenomenon used for time measurement. Astronomy knows of a stellar time, solar time, local time, zone time, nuclear time, etc. The Sun, as well as all other heavenly bodies, partici pates in the motion of the firmament. Except for diurnal motion, the Sun has socalled annual motion, and the totality of the annual motion of the Sun on the firmament is an Ecliptic. If, for example, we fix the location of the constel lations at some definite moment of time and then repeat this observation each month, we can see different pictures of the palate. A view of the starry sky changes continuously: each season has its own picture of vesper constella tions recurring yearly. Therefore, on the expiration of one year the Sun re turns to its initial location in the starry sky. To be oriented conveniently in the stellar world, astronomers divided the entire firmament into 88 constellations. Each of them has its own name. Among the 88 constellations, those that are on the ecliptic, play a special role. These constellations have a generalized title – “the zodiac” (from the Greek word “zoop” – animal), and are widely known throughout the world as symbols (signs) with a variety of allegorical meanings and calendar systems. During its apparent movement along the ecliptic, the Sun intersects 13 constellations. However, astronomers found it necessary to divide the Sun’s Chapter 3 Regular Polyhedrons 149 path into 12 parts (not 13), by uniting the constellations of Scorpions and the Dragon in one unified constellation under the common name Scorpio. The special science of Chronology studies the problems of time measure ment. This science underlies all calendar systems constructed by mankind. In antiquity the creation of calendars was one of the major problems of astronomy. What are “calendars” and “calendar systems?” The word “calendar” orig inated from the Latin word “calendarium” which meant a “debt book”; in An cient Rome interest was tracked in such books and was to be paid on the first day of each month, called “calendas.” The calendar makers paid great attention to periodicities in the motion of the Sun, the Moon, as well as, Jupiter and Saturn, the two gigantic planets of the Solar system. The idea of Jupiter’s calendar with the celestial symbolism of a 12 year animal cycle may be connected to the rotation of Jupiter, which makes its full rotation around the Sun in approximately 12 years (11.862 years). On the other hand, Saturn, the second gigantic planet of the Solar system, makes its full rotation around the Sun in approximately 30 years (29.458 years). By wishing to coordinate the cyclic movements of the gigantic planets, the ancient Chinese in troduced the idea of a 60year cycle of the Solar system. During this cycle, Saturn makes 2 full rotations around the Sun, and Jupiter makes 5 rotations. From antiquity the calendar makers of Eastern and SouthEast Asia made use of the following astronomical phenomena: alternation of day and night, change of lunar phases and alternation of seasons. The use of different astro nomical phenomena resulted in the creation of three kinds of calendars: the lunar one based on the motion of the Moon, the solar one based on the motion of the Sun, and the combination lunarsolar one. 3.3.2. Structure of the Egyptian Calendar The Egyptian calendar that was produced in the 4th millennium BC was one of the first solar calendars. One year in this calendar consisted of 365 days. One Year was divided into 12 months, each Month consisted of 30 days; at the end of the year the 5 holidays that were not part of the month structure were added. Thus, the Egyptian calendar year had the following structure: 365 = 12×30 + 5. It is important to note that the Egyptian calendar is a prede cessor of the contemporary calendar. Why the Egyptians divided the calendar year into 12 months? We know that there were calendars with other numbers of months in the year. For ex ample, in the Maya calendar one year consisted of 18 months, each month being 20 days. Furthermore, why did each month in the Egyptian calendar Alexey Stakhov THE MATHEMATICS OF HARMONY 150 have exactly 30 days? One may ask questions concerning the Egyptian sys tem of time measurement, in particular, regarding the choice of such units of time, such as Hour, Minute and Second. In particular, the question is: why was unit of hour chosen so that 1 day=24(2×12) hours? Further, why 1 hour=60 minutes, and 1 minute=60 seconds? The same questions concerning the choice of the measurement units of angular magnitudes, in particular, why is a cir cumference divided into 360° that is, why 2π=360°=12×30°? We can ask oth er questions, for example: why do astronomers find it expedient to introduce 12 “zodiacal” constellations, instead of the 13 the Sun appears to intersect during its motion along the ecliptic? And why does the Babylonian number system have a rather esoteric radix, the number 60? 3.3.3. Connection of the Egyptian Calendar with the Numerical Characteristics of the Dodecahedron Analyzing the above questions we discover with surprising consistency the following four numbers are repeated: 12, 30, 60 and the derivative num ber 360 (360=12×30). Is there some scientific fact that could give a simple and logical explanation of the use of these numbers in the Egyptian calen dar, and their system of time and angle measurement? To answer this ques tion, we return once again to the regular dodecahedron based on the golden section (Fig. 3.4e). Did the Egyptians know the dodecahedron? Historians of mathematics recognize that the ancient Egyptians had information about the regular poly hedra. The ancient Greek mathematician Proclus attributes to Pythagoras the construction of all 5 regular polyhedra. However, we know that Pythago ras borrowed from the ancient Egyptians many mathematical theorems and discoveries, in particular, the Pythagorean Theorem. Some accounts claim Pythagoras spent 22 years in Egypt and 12 years in Babylon. Therefore, it is possible that Pythagoras could have also borrowed the knowledge about the regular polyhedra from the ancient Egyptians. However, there is more substantial proof that the Egyptians possessed information about all 5 regular polyhedra. In particular, the British Muse um has preserved the dice from Ptolemy’s epoch that have the form of an icosahedron, the Platonic dual to the dodecahedron. These facts allow us to put forward the hypothesis that the dodecahedron was known to the Egyp tians. The very unusual theory of the origin of the Egyptian calendar, as well as the Egyptian measurement system of the time and geometric angles follows from this hypothesis. Chapter 3 Regular Polyhedrons 151 Earlier we found that the dodecahedron has 12 faces, 30 edges and 60 planar angles on its surface. If we accept the hypothesis that the ancient Egyp tians knew the dodecahedron and its numerical parameters 12, 30 and 60, then the scientists of antiquity should not have been surprised, when they discov ered that the cycles of the Solar system are expressed by the same numbers (12year cycle of Jupiter, 30year cycle of Saturn, and 60year cycle of the Solar system). Thus, there is a deep mathematical connection between the Solar system and this perfect spatial figure, the dodecahedron. Scientist of antiquity apparently came to this conclusion. This may explain why the Egyp tians (and Plato) chose the dodecahedron as the “Main Geometric Figure,” that symbolizes the “Harmony of the Universe.” It appears that the Egyptians made all their main systems (calendar system, systems of time and angle mea surement) correspond to the numerical parameters of the dodecahedron! Ac cording to ancient thought, the motion of the Sun on the ecliptic was strictly circular. By then choosing the 12 Zodiac constellations with the distance of 30°, the Egyptians were able to coordinate the yearly motion of the Sun on the ecliptic with the structure of their calendar year: one month correspond ed to the apparent motion of the Sun along the ecliptic between two adja cent Zodiacal constellations! Moreover, the movement of the Sun one de gree along the ecliptic corresponded to one day in the Egyptian calendar! Thus, the ecliptic was divided automatically into 360°. By dividing one day into two parts, the Egyptians thereby automatically divided each half of one day into 12 parts (12 faces of the dodecahedron) and introduced the Hour, a major unit of time. By dividing one hour into 60 minutes (60 planar angles on the surface of the dodecahedron), the Egyptians introduced the Minute, the next important unit of a time. And of course this allowed them to introduce the Second (1 minute = 60 seconds), the smallest unit of time in that period. Thus, by choosing the dodecahedron as the Main Harmonic Figure of the Universe and by following strictly to its numerical characteristics (12, 30 and 60), the Egyptians designed a perfect calendar together with the systems of time and angle measurement that have stood the test over several millen nia. These systems of course correspond to the golden mean “Theory of Har mony,” the underlying proportional basis of the dodecahedron. These surprising conclusions follow from a simple comparison of the dodeca hedron with the Solar system. And if our hypothesis is correct (let somebody attempt to deny it), it follows that for several millennia mankind has lived under the standard of the golden section! And each time, when we look at the index dial of our watch based on the numerical parameters of the dodecahedron 12, 30 and 60, we touch the “Main Secret of the Universe,” the “Golden Section!” Alexey Stakhov THE MATHEMATICS OF HARMONY 152 3.3.4. The Mayan Calendar The Maya were a very advanced race that made great achievements in the fields of mathematics and astronomy. Their calendar was, at the time, the most accurate calendar in the world, surpassed later only by the current Gregorian calendar (introduced in the 16th century). Their calendar was the product of the obsession with measuring long periods of time, which they tracked by means of highly accurate observations of the stars and the planets. The Mayan calen dar had the following structure: 1 year=360+5=20×18+5. Unlike the Egyptian calendar, the Mayan calendar year was divided into 18 months of 20 days. It is clear that the structure of the Mayan calendar was similar to the structure of the Egyptian solar calendar: 1 year=360+5=12×30+5. Remember that the num bers 12 and 30 are numerical parameters of the dodecahedron. However, what does the number 20 refer to in the Maya calendar? Let us address again the icosahedron and the dodecahedron. In these “sacred” figures, there is one more “sacred” numerical parameter: the number of the icosahedron faces is 20 and the number of the dodecahedron vertices is 20! Thus, the Maya undoubtedly used these numerical characteristics of the icosahedron and dodecahedron in their calendar by means of the division of one year into 20 months. 3.4. A DodecahedronIcosahedron Doctrine 3.4.1. Sources of the Doctrine Plato’s cosmology became a source of the socalled DodecahedronIcosa hedron Doctrine, which by a golden thread passes through all human science. The essence of this doctrine consists in the fact that the dodecahedron and the icosahedron are typical forms of Nature in all its manifestations from the macrocosm down to the microcosm. According to the remark of one commentator on Plato’s works, for Plato “all cosmic proportionality is based on the principle of the “golden” or har monic proportion.” Plato’s cosmology is based on the Platonic Solids. Each Platonic Solid symbolized one of the five “beginnings” or “elements”: the Tet rahedron – the body of Fire, the Octahedron – the body of Air, the Hexahedron (Cube) – the body of Earth, the Icosahedron – the body of Water, the Dodeca hedron – the body of the Universe. A representation of the general harmony of the Universe was invariably associated with its embodiment in these regular Chapter 3 Regular Polyhedrons 153 polyhedra. The fact that the dodecahedron, the “Main Cosmic Figure,” was based on the golden section gave to the latter a deep sense in its qualifying to be the “Main Proportion of the Universe.” We mentioned above that, according to Proclus, the commentator on Eu clid’s Elements, Euclid did consider a theory of geometric construction of the Platonic Solids, the “Main Geometric Figures of the Universe,” as the main goal of Euclid’s Elements. Therefore, Euclid placed this major mathematical information in the final, that is, 13th book of his Elements. 3.4.2. A Shape of the Earth The question about Earth’s shape has perennially attracted the attention of scientists since antiquity. When the hypothesis about the spherical shape of the Earth received scientific substantiation, the idea arose that the Earth is a dodeca hedron in shape. Socrates wrote: “The Earth, if to look at it from outside, is sim ilar to the ball consisting of 12 pieces of skin.” Socrates’ hypothesis found further scientific development in the works of physicists, mathematicians and geologists. So, the French geologist de Bimon and the French mathematician Poincare′ pro posed the hypothesis that the Earth is a deformed dodecahedron in shape. In the first half of the 20th century, the Russian geologist Kislitsin pro posed the hypothesis that 400500 million years ago the geosphere’s dodeca hedral shape was converted into the geoicosahedron. However, such trans formation appeared incomplete. As a result, the geododecahedron appeared as if to be inserted into the frame of the icosahedron. Recently, the Moscow researchers Makarov and Morozov have proposed another interesting hypothesis that concerns the shape of the Earth. They supposed that the kernel of the Earth has the shape and properties of a grow ing crystal that is either a dodecahedron or icosahedron. This crystal influ ences the development of all natural processes that occur on our planet. Its energetic field is the cause of this dodecahedralicosahedral structure of the Earth. The influence of this energetic field reveals itself in the fact that we find evidence on the Earth’s surface of the dodecahedron and the icosahedron as if they were inserted into the globe. In recent years, the hypothesis about the icosahedrondodecahedron shape of the Earth was subjected to verification. For this purpose, scientists com bined the axis of the dodecahedron with the axis of the Globe; then they start ed to rotate the dodecahedron around this axis. They saw that the edges of the dodecahedron coincide with the gigantic disturbances of Earth’s crust. Then they started to rotate the icosahedron around the Globe. They saw that Alexey Stakhov THE MATHEMATICS OF HARMONY 154 its edges coincide with the smallersized partitioning of Earth’s crust (moun tain ranges, breaks etc.). These observations support the hypothesis of the closeness of the tectonic framework of Earth’s crust with the shapes of the dodecahedron and the icosahedron. It is as if the vertices of the hypothetical geocrystal are the centers of cer tain anomalies on the planet: in them all global centers of extreme atmospheric pressure including the regions of hurricane origin are located. In one of the hubs of the geoicosahedron (in Gabon) the natural nuclear reactor, which acted 1.7 million years ago, was found. The gigantic mineral fields (for example, Tyumen’s oil field), the anomalies of the animal world (the lake of Baikal), the centers of the development of mankind’s civilizations (Ancient Egypt, Northern Mongo lia, etc.) coincide with many of the vertices of the two polyhedra. All of these examples tend to confirm Plato’s surprising intuition. 3.5. Johannes Kepler: from “Mysterium” to “Harmony” 3.5.1. “Mysterium Cosmographicum” Among the fathers of the new European science there was not a person more mysterious than Johannes Kepler: it seems that he connected two scien tific epochs not only by his “elliptic” astronomical works, but also his unique personality. On the one hand, Kepler was a professional astrologer, a dreamer and visionary, whose style of thinking was unacceptable to the creators of clas sical science, including Galileo and Newton. On the other hand, this astrolo ger, almost medieval in his style of thinking, introduced basic concepts into science. The modern, that is, mechanistic concept of Force was introduced by Kepler. He introduced the concept of Inertia that distinguishes modern phys ics from all former physics. At the same time he introduced the concept of Energy. However, the discovery of New Quantitative Laws of Astronomy was his main scientific achievement. Kepler is the founder of physics of the heav ens. This remarkable wordcombination is a subtitle of his basic work: “New Astronomy based on the Causes, or Physics of the Heavens.” Johannes Kepler is known by all of educated mankind as the author of three famous astronomical laws that overturned astronomical ideas that had existed from antiquity. However, it is less wellknown that these laws were obtained by Kepler as insights resulting from his ambitious research program into Universal Harmony that he pursued at a young age. Chapter 3 Regular Polyhedrons 155 Johannes Kepler was born in 1571 into a poor Protestant family. In 1591, he enrolled in the Tubingen Academy where he received quite good mathe matical education. In the Tubingen Academy the future great astronomer be came acquainted with the heliocentric system of Nicolaus Copernicus. After graduation from the Academy, Kepler obtained a Masters degree and then was appointed mathematics professor at the Graz secondary school (Austria). The small book with the intriguing title Mysterium Cosmographicum, his first astronomical work, was published by Kepler in 1596 at the age of 25. Reading the Mysterium Cosmographicum, it is impos sible not to be surprised by his visionary insights. A deep belief in the existence of the Harmony of the Universe was Kepler’s main idea. Kepler formulated the purpose of his research in the Foreword as follows: “To you kind reader! In this book I intend to demonstrate that our almighty God at the creation of our moving world and at the disposition of the celestial orbits used the five reg Johannes Kepler ular polyhedra that are from Pythagoras’ and Plato’s times (15711630) and up to now have received great honor. He chose the number and proportions of the celestial orbits and also the relations between the planetary motions pursu ant to the nature of the regular polyhedra. I am especially interested in the nature of three things: why are the planets arranged this way and not otherwise, namely, the number, sizes and motions of the celestial orbits.” So, already in the Foreword of his first book the 25yearold Kepler put forward the main problem of contemporary physics, the problem of the natural causes of physical phenomena. Though natural today, this problem in Kepler’s times sounded unusual. In Ptolemy’s and even in Copernicus’ astronomy, this problem had not been formulated. According to that old tradition, the astrono mers considered a problem of their science only in terms of a precise description of planetary motion and the possible prediction of celestial phenomena. 3.5.2. Kepler’s Cosmic Cup How did Kepler answer the surprising questions raised by him in Mysteri um Cosmographicum? After the verification of the numerous hypotheses con nected with the arrangement of planets, Kepler found the following geometri cal model of the Solar system based on the “Platonic solids” (see Fig. 3.11): “Earth’s orbit is the measure of all orbits. We place the dodecahedron around this orbit. The orbit around the dodecahedron is Mars’ orbit. We place Alexey Stakhov THE MATHEMATICS OF HARMONY 156 the tetrahedron around Mars’ orbit. The orbit around the tetrahedron is Jupi ter’s orbit. We place the cube around Jupiter’s orbit. The orbit around the cube is Saturn’s orbit. Then we insert the icosahedron inside Earth’s orbit. The orbit inside the icosahedron is Venus’ orbit. We insert the octahedron inside Venus’ orbit. The orbit inside the octahedron is Mercury’s orbit.” This model gave Kepler an opportunity to collect together all Platonic Solids so that they could unite the then “known” planetary spheres. There are only 5 Platonic Solids and only 5 interplanetary spaces, and all of them appear to allow the regular polyhedra to be placed in them. Could this be mere coincidence? Kepler’s Cosmic Cup (Fig. 3.11) that inserts the Platonic Solids into the crystalline spheres embod ies this model of reality. The most precious trea sure of ancient geometry, the Platonic Solids, were used in Kepler’s astronomy. After that Kepler had the right to say that he comprehended the Universe as if he created it with his own hands. Kepler sent his Mysterium Cosmographicum to Galileo and Brahe and received reassuring respons es from them. Armed with this support, Kepler ad dressed the court of Wurttemberg’s duke Freder Figure 3.11. Kepler’s Cosmic Cup as a model of the Solar ick in hopes of receiving the means for the creation system of the new model of the Universe, the Cosmic Cup, in silver. On his application the duke recommended that he first make it in cop per. The astronomer began to glue a paper model and then threw out the result after one week of hard work. Certainly, the creation of the Cosmic Cup was a great success for the young astronomer. This scientific result had brought scientific renown to Kepler. How ever, this success was incomplete and even somewhat doubtful, because Kepler’s basic scientific purpose had not been achieved. In his first scientific work, Kepler merely anticipated the secret of the world. The Cosmic Cup provided him with access into the invaluable observational data collected in the Heavenly Castle of Tycho Brahe. Kepler assisted the great astronomer in the study of Mars. By using Brahe’s data, the most exact in the world, he intended to polish his new Cosmos, faceted by Platonic Solids, into a cosmic brilliance. It was necessary only to deter mine how the planetary orbits could be placed into the Cosmic Cup. The Cosmic Cup resulted in an important conclusion that disclosed the main secret of the Universe: the Universe is created on the basis of a general geomet ric principle, the regular polyhedra! Unfortunately, Kepler’s joy was premature. In spite of his overfervid character, Kepler had all the characteristics of a seri Chapter 3 Regular Polyhedrons 157 ous scientist. He believed that the theory should fit the observational data. By restraining his delight, Kepler undertook to verify his model. The universal geometric principle allowed Kepler to give answers to two of the three problems that had been raised by him: (1) to explain the number of the then known planets (with the help of the five Platonic Solids, it is possible to construct 6 orbits; the conclusion follows regarding the existence of 6 planets known in that period); (2) to give an answer about the distances between plan ets. The answer to the third problem (about the motion of planets) turned out to be the most difficult and Kepler only attained this answer many years later. 3.5.3. Discovery of Kepler’s First Two Astronomical Laws Kepler’s model in Fig. 3.11 was based upon the supposition that nature’s planetary motion is spherical. By obtaining data from the perennial observa tions of the famous astronomer Brahe, and then carrying out his own observa tions, Kepler decided to reject the astronomical models of his forerunners, Ptolemy and Copernicus, and even his own models. After a thorough study of the planetary orbits, he came to the following conclusion: “The fact that planetary motions are circular is confirmed by their inces sant recurrence. The human intellect, by extracting this truth from experi ence, at once concludes from here, that the planets are rotated on the ‘ideal’ circles, because among the planar figures a circle and among the spatial fig ures a sphere can be considered as the most perfect geometric figures. Howev er, after closer examination we can conclude that experience gives another result, namely, the planetary orbits differ from simple circles.” The results of this work are presented in Kepler’s main book A New As tronomy Based on the Causes, or Physics of the Heavens published in 1606. The importance of this book consists, first of all, of the fact that Kepler therein formulates the first two of three astronomical laws named after him. Accord ing to Kepler’s First Law, the orbit of each planet is an Ellipse with the Sun as one of its foci. Kepler’s Second Law asserts that the areas, covered by the radii vectors of the planets, are constant. Kepler’s Laws are the first quantitative laws that are of great importance for the development of astronomy. By testing the Cosmic Cup, Kepler found that this model based upon circular planetary orbits did not fit the experi mental data. Therefore, Kepler decided to reject the idea of circular planetary motion. After numerous and difficult calculations Kepler found that the or bits of planets were not circles but Ellipses. This result dramatically altered much of the basis of previous cosmology. Alexey Stakhov THE MATHEMATICS OF HARMONY 158 3.5.4. “Harmonices Mundi” Kepler’s third astronomical law, formulated by him in Harmonices Mundi (The Harmony of the World) completed the creation of a new astronomy. Kepler’s first two laws could not answer one major question of astronomy. How is it that the planetary distances from the Sun change? Kepler tried to grasp a new principle for the solution to this challenge. He used his philo sophical views going back to Pythagoras and Plato. According to Kepler’s deep belief, Nature was created by God on the basis of not only mathematical, but also harmonious principles. Kepler believed in the “music of the spheres” that generate bewitching melodies embodied not in sounds, but in the move ments of the planets capable of generating harmonious chords. Based on this idea, Kepler combined mathematical and musical arguments to discover the third law of planetary motion that asserts the following: “If Т is the cycle of time of a planet’s orbit around the Sun, and D is its middle distance from the Sun, then they are connected by the following cor relation: T 2=kD3, where k is a constant value equal for all planets.” Kepler arrived at the following conclusion from this discovery: “Thus, the heavenly movements are a never ending polyphonic music that is perceived by the human intellect, but not the ear.” The Russian scientist Predtechensky (18601904), Kepler’s biographer, wrote about this Kepler’s discovery as follows: “The wonderful harmony, reign ing in the world, was perceived by Kepler not only in an abstract sense of the organization of the Universe. The harmony was sounding in its poetic soul by a true form of music, which could be understood by us if we could enter into the circle of his ideas and would be imbued with his mighty enthusiasm about the marvelous construction of the world and the Pythagorean belief in numerical relations. Really, it is surprising that what makes sounds ‘beautiful’ for hearing depends upon a strict numerical ratio, for example, between the lengths of the strings, which produce the sounds as discovered by Pythagoras. But in Kepler, undoubtedly, part of Pythagoras’ soul lived on, and it is not accidental that he saw the numerical ratios in the planetary cosmos.” Kepler’s third Law is the outstanding scientific result, at the summit of his career. The result obtained filled Kepler’s soul with great joy and grati tude for the Creator. He expressed this gratitude in the following words: “The wisdom of the Creator is endless and his glory and power are bound less. You, the heavens, glorify and praise be to Him! The Sun, the Moon and the planets, glorify God with their ineffable language! Celestial harmonies, comprehended by His wonderful creations, do glorify and praise to Him! Let Chapter 3 Regular Polyhedrons 159 my soul eulogize the Creator! Everything is built by Him and everything is embodied within Him. Everything the best which is known to us is created by Him in our bustling life. Praise, Honor and Glory to Him for all of eternity!” The dramatic period in astronomical development was completed. This pe riod ended with the discovery of three major astronomical laws of planetary motion, Kepler’s Laws. This history started with an original model of the Solar system based on the Platonic Solids. Though Kepler’s model, in the end, ap peared somewhat erroneous, Kepler did not reject from this model what he de scribed in his Mysterium Cosmographicum. He always considered this model to be one of his greatest scientific achievements. By submitting to the requests of his friends, Kepler in the twilight of his life decided to undertake the second main issue of his first book “for benefit not only to booksellers, but also scien tists.” Addressing the new readers, Kepler wrote in the dedication with pride: “Almost 25 years have passed since I issued a small book Mysterium Cos mographicum. Though in that time I was very young and this publication was my first astronomical work, nevertheless the success accompanying this book in subsequent years, testifies eloquently that no one could write a first book with a more substantive, successful and valuable treatment of the subject. It is as if an oracle from the heavens dictated through me the chapters of this book, because all of them, as is generally accepted, were excellent and corresponded to the truth. During the 25 years, the chapters of this book illuminated my way in astronomy many times. Almost all my astronomical works, which I published by this time, had their beginning in one or another chapter of my first book, and therefore these later books can be considered to be a more in depth or more complete presentation of these original chapters.” 3.5.5. Life through the Centuries Kepler’s life is an example of scientific selflessness based upon the timeless belief in the Harmony of the Universe. His entire life was a struggle on two fronts. On the one hand, the “mathematician of the famous province Stiria” strug gled against poverty and the almost intolerable “life” of the poor having so many children. His life was accompanied by the depression of his wife, the untimely deaths of his children, the charges of sorcery against his mother, and the dull ness of coreligionists. On the other hand, Kepler’s scientific life was a world of calculations, improbable in their complexity. “The intense, incessant and vigor ous thoughts” were the basis of Kepler’s unique scientific life. However, old age was coming. Kepler’s death (in 1630) interrupted his work on the last book Somnium (Dreams), a science fiction novel about flight Alexey Stakhov THE MATHEMATICS OF HARMONY 160 to the Moon. Unfortunately, the Harmony was not written in this book. There were no captious verifications; there were no new hypotheses. Kepler was tired: “My brain got tired when I attempted to understand what I wrote, and it is already difficult for me to reestablish a connection between figures and text found by me at an earlier time …” So, the drama had been finished. With Kepler’s death, his discoveries were forgotten. Even the wise Descartes did not know about Kepler’s works. Gali leo had not found necessary to read his books. Only in Newton’s works Ke pler’s Laws find a new life. However, the Harmony is not of interest to New ton. He is interested in Equations. A new era had arrived. Kepler’s life terminated the epoch of “scientific romanticism,” the epoch of the Harmony of the Golden Section that is characteristic of the Renais sance. On the other hand, his scientific works became the beginning of the new science that started with the works of Descartes, Galileo and Newton. And in conclusion we once again remember Kepler’s wellknown state ment about the Golden Section: “Geometry has two great treasures: one is the Theorem of Pythagoras; the other, the division of a line into extreme and mean ratio. The first we may compare to a measure of gold; the second we may name a precious jewel.” For those, who treat skeptically Kepler’s comparison of the Pythagorean Theorem with the Golden Section, we should remind them that Kepler was not only a great astronomer, but also a great mathematician! Predictions of great scientists can sometimes dramatically advance the development of a sci ence. The development of modern science confirms that Kepler was right: the Golden Section becomes one of the major ideas of modern science! Unfortunately, after Kepler’s death, the Golden Section, that was consid ered by him to be one of the great “treasures of geometry,” was forgotten. This unfortunate ignorance was continued for almost two centuries. With few ex ceptions interest in the Golden Section was revived only in the 19th century! 3.6. The Regular Icosahedron as the Main Geometrical Object of Mathematics 3.6.1. Felix Klein Among the five Platonic Solids, the icosahedron and the dodecahedron have a special place. In Plato’s cosmology the icosahedron symbolizes Water, and the dodecahedron the Harmony of the Universe. These two Platonic Chapter 3 Regular Polyhedrons 161 Solids are connected directly with the pentagon and through it the Golden Section. The dodecahedron and icosahedron give rise to the socalled “icosa hedrondodecahedron doctrine” that permeates the history of human culture from Pythagoras and Plato up to the present. It is not an accident that this doctrine received a unique boost or development in the works of the German mathematician Felix Klein (18491925). Felix Klein, a graduate of the University of Bonn, was born in 1849 and died in 1925. Beginning in 1875 he worked as a Professor of the Higher Technical School in Munich, then from 1880 as a Professor of the University of Leipzig. In 1886 he arrived in Gettingen, where he headed the Mathematical In stitute of the University of Gettingen. During the first quar ter of the 20th century this Mathematical Institute was rec ognized as the World’s leading center of mathematics. Klein’s main works were dedicated to NonEuclidean Felix Klein (18491925) geometry, and the theories of continuous groups, algebraic equations, and elliptic and automorphic functions. Klein presented his ideas in the field of geometry in a Comparative Consideration of the New Geometri cal Researches (1872) known under the title Erlangen’s Program. According to Klein, each geometry is some invariant theory of a special group of transformations. By dilating or narrowing down the group, it is pos sible to pass from one type of geometry to another. The Euclidean geometry is a science about invariants of the metric group, projective geometry about invariants of the projective group, etc. The classification of transformation groups gives us the classification of the geometries. A proof for the existence of different NonEuclidean geometries is considered to be Klein’s most essen tial achievement. 3.6.2. Elementary Mathematics from the Point of View of Higher Mathematics Felix Klein was not only a wellknown mathematiciantheorist, but also a reformer of mathematical education in the schools. Prior to World War I he organized the commission on reorganization of mathematical teaching. The book Elementary Mathematics from the Point of View of Higher Mathematics was devoted to the development of a solution to this problem. Mathematics in the 19th century produced a number of remarkable ideas that influenced all branches of knowledge and engineering. The main idea behind the reformers headed by Klein was to increase the role of mathematics Alexey Stakhov THE MATHEMATICS OF HARMONY 162 and natural science in the general education. The study of natural sciences and mathematics, a deepening of the connection between theoretical and ap plied mathematics, the introduction of functional analysis and calculus into mathematical teaching, and a broad use of graphic methods are the main prin ciples suggested by Klein and his followers as a basis for mathematical educa tion. It is important to recognize that Klein’s ideas are rather topical for mod ern mathematical education. 3.6.3. Role of the Icosahedron in Mathematical Progress Besides the Erlangen Program and other outstanding mathematical achievements, Klein’s greatness consists of the fact that 100 years ago he pre dicted an outstanding role for the Platonic Solids, in particular, the icosahe dron in the future development of science. In 1884, Klein published the book Lectures on the Icosahedron and Solution of the 5th Degree Equations [58] ded icated to a geometric theory of the icosahedron. The icosahedron (and its dual, the dodecahedron) play a special role in “living” nature; many viruses and other living things have the shape of the icosahedron, that is, the icosahedron shape and pentagonal symmetry play a fundamental role in the organization of living substance. According to Klein, the tissue of mathematics extends widely and freely by the sheets of the different mathematical theories. However, there are geo metric objects, in which many mathematical theories converge. Their geome try unites these mathematical theories and allows us to embrace the general mathematical sense of the various theories. The icosahedron, in Klein’s opin ion, is just such a mathematical object. Klein treats the regular icosahedron as the central mathematical object, from which the branches of the five mathe matical theories follow, namely: Geometry, Galois’ Theory, Group Theory, The ory of Invariants and Differential Equations. Thus, the great mathematician Felix Klein following after Pythagoras, Pla to, Euclid, Johannes Kepler could realize the fundamental role of the Platonic Solids, in particular the icosahedron, for the development of science and math ematics. Klein’s main idea is extremely simple: “Each unique geometrical object is somehow or other connected to the properties of the regular icosahedron.” Unfortunately, Klein’s contemporaries could not understand and appre ciate the revolutionary importance of Klein’s idea proposed by him in the 19th century. However, its significance was appreciated one century later, when the Israeli scientist Dan Shechtman discovered in 1982 a special alloy with an icosahedral phase called Quasicrystals. And the famous researchers Chapter 3 Regular Polyhedrons 163 Robert F. Curl, Harold W. Kroto and Richard E. Smalley discovered in 1985 a special kind of carbon called Fullerenes. In 1996, they became Nobel Prize winners for their discovery. It is important to emphasize that the quasicrys tals are based on the Platonic icosahedron and the fullerenes on the Archimedean truncated icosahedron. 3.7. Regular Polyhedra in Nature and Science 3.7.1. Symmetry Groups of Regular Polyhedra The influence of Klein’s Erlangen Program on school geometry is especial ly important. The influence of Klein’s “group approach” can be traced in all themes of school geometry. Each geometric figure F determines some group of movements; this group contains all those movements that convert the figure F in a manner similar to itself. This is called the Symmetry Group of the figure F. In many respects, knowledge of the symmetry group of the figure F deter mines the geometric properties of that figure. Let us consider in greater detail some im portant concepts of symmetry theory. We start with the Symmetry Plane P that is familiar to us from the previous Chapter. We offer to the reader to be convinced that a square has four planes of symmetry (4Р), however, a rectangle has only two planes of symmetry (2Р). It can be easily proven that a cube has 9 planes of sym metry (Fig. 3.12), that is, 9Р. Let us now consider the second type of sym metry elements, the Axis of Symmetry. The axis Figure 3.12. The 9 planes of symmetry of the cube of symmetry is based upon a straight line, around which the identical parts of a symmetric figure can be repeated a par ticular number of times. These identical parts are located so that after a turn around, the symmetry axis on the certain angle the figure occupies the same position as before the turn. As a result, the figure comes to “selfcoincidence.” A faceted glass in a glass holder is the best visual example of “selfcoincidence.” Taking the glass out of the glass holder and then inserting it back in the changed position, we execute the operation of “selfcoincidence.” The number of all Alexey Stakhov THE MATHEMATICS OF HARMONY 164 possible “selfcoincidences” that can be carried out around a given axis is named its Symmetry Order. Usually the axis of symmetry is denoted by the capital letter L, but its order is denoted by a small subscript number that stands after this capital letter. Thus, for example, the symmetry axis of the 3rd order is denoted by L3. It is clear, that the equilateral triangle has the symmetry axis L3, the square – L4, the pentagon – L5, and the circle has the symmetry axis of the infinite order L∞. As an example, for demonstration of symmetry axes, we examine a cube (Fig. 3.13). a) b) c) Figure 3.13. The symmetry axes of the cube: 3L4(a), 4L3(b), 6L2(c) We can draw three axes of symmetry of the 4th order through the center of a cube perpendicularly to each pair of opposite faces (Fig. 3.13a). This means, that a cube has 3 axes of symmetry of the fourth order, that is, 3L4. A cube has 8 vertices. We can draw a triple symmetry axis that coincides with the corresponding diagonal of the cube through each pair of the opposite vertices of the cube (Fig. 3.13b). This means, that the cube has 4 symmetry axes of the 3rd order, that is, 4L3. The cube has 12 edges. We can draw double symmetry axes through the middles of each pair of the opposite edges, in parallel to the diagonals of the faces (Fig. 3.13с). This means, that the cube has 6 symmetry axes of the 2nd order, that is, 6L2. Therefore, the full set of the cube axes of symmetry is the following: 3L44L36L2. There are geometric figures that have symmetry axes of the infinite order L∞ . The socalled “figures of rotation,” cylinder, cone, etc. have similar axes. Any diameter of a fullsphere is the axis of the type L∞. This means, that a fullsphere has an infinite set of symmetry axes of infinite order, that is, ∞ L∞. Now, let us consider one more element of symmetry, the Center of Symmetry. The center of symmetry is a special point inside the figure, when any straight line, drawn through this point, meets an identical point of the figure at equal distances from this point. The cube, for example, has a center of symmetry. Usually for the description of symmetry of some geometric figure a full set of symmetries is used. For example, the symmetry group of a snowflake is L66P. This means that a snowflake has one symmetry axis of the 6th order L6, that is, Chapter 3 Regular Polyhedrons 165 the snowflake “selfcoincides” 6 times at its rotation around the symmetry axis, and 6 planes of symmetry. The symmetry group of camomile having 24 petals is L2424P, that is, the camomile has one symmetry axis of the 24th order and 24 planes of symmetry. By uniting all symmetry elements of the cube, we can write the following symmetry group of the cube: 3L44L36L29PC. The cube Table 3.5. Symmetry groups of the Platonic Solids is, of course, Polyhedron A shape of faces Symmetry group one of 5 Pla 4 L3 3 L2 6 P Tetrahedron Equilateral triangles tonic Solids. 3 L4 4 L3 6 L2 9 PC Cube Squares Table 3.5 pre sents the sym 3 L4 4 L3 6 L2 9 PC Octahedron Equilateral triangles metry groups 6 L5 10 L3 15 L2 15 PC Dodecahedron Pentagons of all Platonic 6 L5 10 L3 15 L2 15 PC Icosahedron Equilateral triangles Solids. Analysis of the symmetry groups of Platonic Solids, given in Table 3.5, shows that the symmetry groups of the cube and the octahedron, on the one hand, and the dodecahedron and the icosahedron, on the other hand, coin cide. Thus the dodecahedron is dual to the icosahedron, and the cube is dual to the octahedron. 3.7.2. Applications of Regular Polyhedra in the Living Nature Ernst Heinrich Philipp August Haeckel (18341919) was the eminent Ger man biologist and philosopher. He became famous following the publication of his book Kunstformen der Nature (Artshapes of Nature) Haeckel wrote in his book: “Nature has an inexhaustible number of sur prising creations, which by beauty and variety far surpass all works created by human art.” Nature’s creations presented in Haeckel’s book are beautiful and sym metric (see Fig. 3.14). This fact is an inseparable property of natural harmo ny. In his book he gave the examples of the singlecell organisms similar to the icosahedron in shape. The icosahedron attracted the attention of biolo gists in their disputes concerning the shape of viruses. The virus cannot be absolutely round as was considered earlier. To establish its form, the biolo gists took the various polyhedra and directed a light on them under the same angle, such as a stream of atoms on the virus. It is proved that only the icosa hedron gives a precisely identical shadow. It is considered that the geomet rical properties of the icosahedron allow it to preserve genetic information in the best possible manner. Alexey Stakhov THE MATHEMATICS OF HARMONY 166 Figure 3.14. Artshapes of Nature from Haeckel’s book Kunstformen der Nature (1904) The regular polyhedra are the most “preferred” and “symmetric” figures. Nature makes wide use of this property of the polyhedra. The crystals of many chemical substances have the shape of regular polyhedra. For example, a crys tal of table salt NaCl has the shape of a cube, a monocrystal of the potassium alum has the shape of an octahedron, a crystal of the chemical substance FeS has the shape of a dodecahedron, a sodium sulphite – a tetrahedron, a boron Br – an icosahedron and so on. Chapter 3 Regular Polyhedrons 167 Today it has been proven that the process of the formation of a human embryo from the ovum is carried out by its divisions according to the “bina ry” law, that is, at first the ovum is transformed into two cells, these new cells are transformed in turn into four cells, etc. Earlier we considered that at the second stage of the division the four cells form a square. In fact, it occurs in another manner: at the second stage of the division, the four cells form the tetrahedron (Fig. 3.4а). The cells are further divided into eight; together forming a geometric figure that consists of one tetrahedron point ing upward and the oth er tetrahedron pointing downwards, connected together. As a result we obtain the Star Tetrahe dron that reminds one of Egg of life Tetrahedron Star Tetrahedron the Egg of Life in esoter Figure 3.15. A star tetrahedron ic sciences (Fig. 3.15). 3.7.3. “Parquet Problem” and Penrose Tiling Since ancient times, “parquet’s problem” in geometry has been the prob lem of how to fill a plane with regular polygons. Already the Pythagoreans proved that only regular (equilateral) triangles (symmetry axis of the 3rd or der), squares (symmetry axis of the 4th order) and hexagons (symmetry axis of the 6th order) can be solutions to this problem. “Parquet’s problem” has a direct relation to the main law of crystallography, according to which only the symmetry axes of the 3rd, 4th and 6th 1 order are allowed in crystals. The symme 1/τ 36° try axes of the 5th order and more than 6 are prohibited in crystals. The English mathematician Sir Roger а) Penrose was one of the first scientists, who 1 found another solution to “parquet’s prob 72° lem.” In 1972, he covered a plane in a non τ periodic manner, by using only two simple polygons. In the simplest form, the Penrose’s Tiles are a nonrandom set of rhombi of two types, the first one (Fig. 3.16a) has the in b) ternal angle 36°, and the second one (Fig. Figure 3.16. Penrose tiles: (a) “thin” 3.16b) has the internal angle 72°. rhombus; (b) “thick” rhombus Alexey Stakhov THE MATHEMATICS OF HARMONY 168 To understand the mathematical essence of the Penrose Tiles, we return to the Pentagon and Pentagram (Fig. 3.17). The pentagram contains a number of char A acteristic isosceles triangles. The triangle of the type ADC is the first of them. The acute angle at the vertex of A is 36°, and the ratio of the side F G B AC=AD to the base DC is equal to the golden E mean, that is, the given triangle is the golden isos H L celes triangle. Note that the pentagram consists of 5 identical golden isosceles triangles ADC, K BED, CAE, DBA, and ECB (the 5 crossing trian C D gles). Besides, we have in the pentagram Figure 3.17. Pentagram 5 smaller “golden” isosceles triangles AGF, BHG, CKH, DLK, and EFL similar to ACD. If we now take two such triangles and connect them together by their bases, we obtain the Penrose’s tile displayed in Fig. 3.16a and named “Thin” Rhombus. The “thin” rhombus has four verti ces with the following angles: 36°, 36°, 144°, 144°. Now, let us consider one more type of isosceles triangle presented in the pentagram, for example, ABE. In such triangle the acute angles at the vertices Е and B are each 36°, and the obtuse angle at the vertex A is 108°. Note that the ratio of the base ЕВ of the triangle ABE to its sides AE=AB is equal to the golden mean, that is, this triangle is also the golden isosceles triangle. Note also, the pentagram has 5 identical golden isosceles triangles of this kind, name ly, ABE, BCA, CDB, DEC, and EAD. If we connect two such triangles together at their bases, we obtain the second Penrose tile represented in Fig. 3.16b and named “Thick” Rhombus. The thick rhombus has the four vertices with the angles: 72°, 72°, 108°, 108°. Below in Fig. 3.18 we can see a process of sequential construction of the Penrose Tiling. Take 5 “thick” rhombi and connect them together, as shown in Fig. 3.18a. Then, we add to the figure in Fig. 3.18a the “thick” and “thin” rhombi, as shown in Fig. 3.18b. Figure 3.18c is a further development of the Penrose tiling. It is proved that the ra tio of the number of “thick” rhombi (Fig. 3.16b) to the number of “thin” rhombi (Fig. 3.16a) in the Penrose a) b) c) tiling aims in the limit for Figure 3.18. Penrose tiling the golden mean. Chapter 3 Regular Polyhedrons 169 3.7.4. Quasicrystals On November 12, 1984 in a small article, published in the authoritative journal Physical Review Letters, the ex perimental proof of the existence of a metal alloy with ex clusive physical properties was presented. The Israeli physi cist Dan Shechtman was the author of this article. Dan Shechtman is Philip Tobias Professor of Materials Science at the Israel University Technion. A special alloy discovered by Professor Shechtman in 1982 and called Quasi The Israel physicist Dan Shechtman crystal is the focus of his research. By using methods of elec tronic diffraction, Shechtman found new metallic alloys having all the symptoms of crystals. Their diffraction pictures were composed from the bright and regular ly located points similar to crystals. However, this picture is characterized by the socalled “icosahedral” or “pentagonal” symmetry, strictly prohibited according to geometric reasons. Such unusual alloys are called “quasicrystals.” The concept of quasicrystals generalizes and completes the definition of a crystal. Gratia wrote in the article [148]: “A concept of the quasicrystals is of fundamental interest, because it extends and completes the definition of the crystal. A theory, based on this concept, replaces the traditional idea about the ‘structural unit,’ repeated periodically, with the key concept of the distant order. This concept resulted in a widening of crystallography and we are only beginning to study the newly uncovered wealth. Its significance in the world of crystals can be put at the same level with the introduction of the irrational to the rational numbers in mathematics.” What are quasicrystals? What are their properties and how we can de scribe them? We mentioned above that according to the Main Law of Crystal lography some strict restrictions are imposed on the structure of a crystal. According to classical ideas, the crystal is constructed from one single cell. The identical cells should cover a plane densely without any gaps. As we know, the dense filling of a plane can be carried out by means of Equi lateral Triangles (Fig. 3.19а), Squares (Fig. 3.19b) and Hexagons (Fig. 3.19d). A dense filling of the plane by means of Pentagons is impossible (Fig. 3.19c). а) b) c) d) Figure 3.19. A dense filling of a plane can be carried out by means of equilateral triangles (а), squares (b) and hexagons (d) Alexey Stakhov THE MATHEMATICS OF HARMONY 170 Penrose tiling (Fig. 3.18) that is a “planar analogy” of quasicrystals was used for the theoretical explanation of the quasicrystal phenomenon. In the spatial model, the “regular icosahedrons” (Fig. 3.20) played the role of “Pen rose rhombi” in the planar model. By using regular icosahedrons, we can ful fill a dense filling of threedimensional space. What is the practical significance of the discovery of quasicrystals? Gratia wrote in [148] that “the mechanical strength of the quasicrystals increased sharply; here the absence of periodicity resulted in slowing down the distri bution of dislocations in comparison to the traditional metals .… This property is of great practical significance: the use of the “icosahedral” phase allows for light and very stable alloys by means of the inclusion of smallsized frag Figure 3.20. A regular icosahedron ments of quasicrystals into the aluminum matrix.” What is the significance of the discovery of quasicrystals from the point of view of the main idea of our book, the golden mean? First of all, this discov ery is a great triumph for the “icosahedrondodecahedron doctrine,” which passes throughout the history of natural sciences and is a source of profound ly practical scientific ideas. Secondly, the quasicrystals shattered the con ventional picture of an insuperable barrier between the mineral world where the “pentagonal” symmetry was prohibited, and the living world, where the “pentagonal” symmetry is widespread. Note that Dan Shechtman published his first article about the quasicrys tals in 1984, that is, exactly 100 years after the publication of Felix Klein’s Lectures on the Icosahedron in 1884. This means that this discovery is a wor thy gift to the centennial anniversary of Klein’s book, in which the famous German mathematician predicted an outstanding role for the icosahedron in future scientific development. 3.7.5. Fullerenes Fullerenes were an important modern discovery in chemistry. This discovery was made in 1985, several years after the quasicrys tal discovery. The “fullerene” is named after Buckminster Fuller a) b) (1895 1983), the American de Figure 3.21. Fuller’s inventions: signer, architect, poet, and inven (a) Buckminster Fuller; (b) Fuller’s geodesic dome Chapter 3 Regular Polyhedrons 171 tor. Fuller created a large number of inventions, primarily in the fields of design and architecture. The bestknown invention was the Geodesic Dome based on the truncated regular icosahedron (Fig. 3.21b). The geodesic dome, Montreal Biosphere, de signed by Buckminster Fuller for the American Pavilion at Expo 67 (Fig. 3.22), is Fuller’s best known architectural construction. Figure 3.22. The Montreal Fuller was the author of many inventions in Biosphere, formerly the American Pavilion of Expo 67 designing and architecture. In the USA, a post age stamp was produced immortalizing Buckminster Fuller’s contributions to architecture and science (Fig. 3.23). Fuller’s head on this postage stamp de picts the structure of his geodesic dome. After the discovery of Fullerenes, the name of Buckminster Fuller became famous worldwide. The title “fullerenes” refers to the carbon molecules С60, С70, С76, and С84 in which all the atoms are on a spherical or spheroid surface. In these molecules the atoms of carbon are located at the vertices of regular hexa gons or pentagons that cover the surface of the sphere or spher oid. We start from a brief history of the C60 molecule. This mol ecule plays a special role among the fullerenes. It is character Figure 3.23. The U.S. postage stamp ized by the greatest symmetry and as a consequence is highly immortalizing stable. By its shape, the С60 molecule reminds one of a typical Buckminster Fuller white and black soccer ball (Fig. 3.24) that has the structure of a truncated regular icosahedron (Fig. 3.5e, Fig. 3.6). The atoms of carbon in this molecule are located on the spherical surface at the vertices of 20 regular hexagons and 12 regular pentagons; here each hexagon is surrounded by three hexagons and three pentagons, and each pentagon is surrounded by five hexagons (Fig. 3.25). The most striking property of the C60 Figure 3.24. The Soccer Ball molecule is its high degree of symmetry. There are 120 symmetry operations that convert the mole cule into itself making it the most symmetric molecule. It is not surprising that the shape of the C60 molecule has attracted the attention of many artists and mathemati Figure 3.25. The cians over the centuries. As mentioned earlier, the truncat structure of the С60 molecule ed icosahedron was already known to Archimedes. The old Alexey Stakhov THE MATHEMATICS OF HARMONY 172 est known image of the truncated icosahedron was found in the Vatican library. This picture was from a book by the painter and mathematician Piero della Francesca. We can find the truncated icosahedron in Luca Pacioli’s Divina Pro portione (1509). Also Johannes Kepler studied the Platonic and Archimedean Solids actually introducing the name “truncated icosahedron” for this shape. The fullerenes, in essence, are “manmade” structures following from funda mental physical research. They were discovered in 1985 by Robert F. Curl, Harold W. Kroto and Richard E. Smalley. The researchers named the newlydiscovered chemical structure of carbon C60 the Buckminsterfullerene in honor of Buckmin ster Fuller. In 1996 they won the Nobel Prize in chemistry for this discovery. Fullerenes possess unusual chemical and physical properties. At high pressure the carbon С60 becomes firm, like diamond. Its molecules form a crystal structure as though consisting of ideally smooth spheres, freely rotating in a cubic lattice. Ow ing to this property, С60 can be used as firm greasing (dry lubricant). The fullerenes also possess unique magnetic and superconducting properties. 3.8. Applications of Regular Polyhedra in Art 3.8.1. Leonardo da Vinci’s Methods of Regular Polyhedra Representation Many authors pay particular attention to Leonardo da Vinci’s original methods of the spatial representation of icosahedron, dodecahedron and trun cated icosahedron, for the book Divina Proportione (1509) by his contempo rary, the Franciscan monk and mathematician Luca Pacioli (1445 1514). It is probably impossible to consider Leonardo’s participation in studying such perfect geometrical figures as Platonic and Archimedean Solids as a mere co incidence. And what is more, this fact is deeply symbolic. A true titan of the Renaissance, artist, sculptor, scientist, engineer and inventor, Leonardo da Vinci (1452 1519) is a symbol of the inseparable bond between Art and Sci ence. His interest in such fine and highly symmetrical figures as regular and semiregular polyhedra was natural. Below (Fig. 3.26) we examine the different representations of the dodeca hedron used by Leonardo da Vinci in Pacioli’s Divina Proportione. Leonardo used two methods, a method of Rigid Edges (Fig. 3.26a) and a method of Con tinuous Faces (Fig. 3.26b). Comparison of these methods with the example of the dodecahedron convincingly shows the advantage of the method of rigid edges. Chapter 3 Regular Polyhedrons 173 The essence of the method of rigid edges consists in the fact that the faces of the polyhedron are represented as “empty,” not continuous. Strictly speaking, the faces are not represented at all; they exist only in our imagination. However, the edges of the polyhedron are represented not by geometrical lines (which have neither width nor thickness), but by the rigid threedimensional segments. Such techniques of polyhedron repre sentation allow the spectator to accurately determine, first of all, which edges belong to the front and which belong to the back fac es of the polyhedron (which is practically impossible when the edges are represented by geomet rical lines). On the other hand, we can look through a geometrical a) b) body to see the body in perspec Figure 3.26. Leonardo’s representations of the tive and depth. It is clear that this dodecahedron by the methods of (а) rigid edges is impossible if we use the method and (b) continuous faces in Pacioli’s Divina Proportione of continuous faces. Below (Fig. 3.27) we see Leonardo’s representation of the truncated icosa hedron by the method of rigid edges. At the top of the picture we see the Latin inscription Ycocedron Abscisus (truncated icosahedron) Vacuus. The Latin word Vacuus meant, that the faces of the polyhedron are represented as “empty,” not continuous. It is necessary to note that the representation of polyhedra by the method of rigid edges began to be used widely in science and works of art following Leonardo. As an example, Johannes Kepler used the method of rigid edges for the representation of the polyhedra from which he constructs his Cosmic Cup (Fig. 3.11). 3.8.2. Pacioli’s Polyhedron As we saw previously, Pacioli was one of the great est mathematicians of 15th century Europe. He also invented the principle of the socalled double record Figure 3.27. Leonardo’s representation of the used now in all modern systems of bookkeeping. That truncated icosahedron by is why, he can be called “the father of modern book the method of rigid edges in keeping.” However, the creative works of Luca Pacio Pacioli’s Divina Proportione Alexey Stakhov THE MATHEMATICS OF HARMONY 174 li, who was a very mysterious person in his time, have up to now been the source of fierce disputes amongst historians of science. It is known that Luca Pacioli was born in 1445 in the Italian city Borgo San Sepulcro. In his childhood he was a pupil of the artist and mathematician Piero della Francesca. Later he was a student at the University of Bologna, which during the 15th and 16th centu ries was one of the best educational institutions in Europe (at various periods Copernicus and Durer were its students). In 1472, Pacioli came back to his native city Borgo San Sepulcro and started to write his best known book Sum ma de Arithmetica, Geometria, Proportioni et Proportionalita. This book was not published in Venice until 1494. In 1496, he was invited to Milan to give the lectures on geometry and mathematics. Here is when he met Leonardo da Vin ci. After reading Pacioli’s Summa, Leonardo stopped writing his own book on geometry and started to illustrate Pacioli’s new book Divina Proportione. Some researchers accuse the author of Div ina Proportione of plagiarism of the unpublished manuscript belonging to Pacioli’s teacher Pie ro Della Francesca. However, it is rather doubt ful, if we take into consideration that in the pe riod when Pacioli was writing this book, he was already a wellknown mathematician, and his glory reverberated throughout all of Italy. Pa cioli is well known to us today owing to the Figure 3.28. Jacopo de Barbari’s portrait (Fig. 3.28) painted of him by the Ital painting of Luca Pacioli ian artist Jacopo de Barbari (1440 1515). Bar bari’s picture is revealing in several respects, first of all, in respect of the presen tation of Luca Pacioli’s personality. Numerous compositional details in Barbari’s picture are full of deep scientific sense. The artist demonstrates an understand ing of the interrelation between Art and Science which was peculiar to Renais sance experts. Pacioli in the robe of the Franciscan monk is represented by standing behind the table of geometrical tools and books (in the lower righthand corner of the picture we can see a model of the dodecahedron). The attention of Pacioli and the handsome young man, who stands on the right and somewhat behind Pacioli, is directed to the glass model of the polyhedron that hangs in the upper left hand corner of the picture. The choice of the poly hedron is not accidental: it is Pacioli’s rhombic cube Figure 3.29. Pacioli’s octahedron (Fig. 3.29). rhombic cubeoctahedron Chapter 3 Regular Polyhedrons 175 3.8.3. Albrecht Durer The identity of the young man, who is near to Pacioli in Jacopo de Bar bari’s picture (Fig. 3.28), has often caused disputes amongst historians of art. Some of them assert that this is Barbari’s selfportrait, others identifying the individual with Albrecht Durer (1471 1528), German painter, draughtsman and art theorist who is generally known as the greatest German Renaissance artist. However, and this is very important in our context, Durer was amazed by the art method of Barbari, who later created his pictures on the basis of a deep study of the system of proportions, that is, strict use of particular ratios of parts of the represented objects among themselves. Durer was one of the first artists to start studying the laws of perspec tive. He dreamed of meeting with the glorified Italian artists to study their art works and to compete with them. With this purpose in mind Durer trav eled to Italy in 1505. We do not know his teachers at the school of perspec tive, but we do know that throughout his life Durer continued to teach at this school. He wrote several books on the theory of arts. One of them, The Painter’s Manual, was dedicated to geometry and perspective and was pub lished in Nuremberg in 1525. His last work on human proportion was pub lished in 1528 following his death. Durer’s books are a serious scientific contribution to the theory of per spective and stereometry of polyhedra. He described several Archimedean Solids unknown at that time, and also developed and published for the first time books that included spatial models of the planar unfolding of various polyhedra, including the unfolding of the truncated icosahedron. Today a similar unfolding of volumetric models of polyhedra is widely used to study the elementary forms of crystals, the structures of molecules (fullerenes, for example), and vi ruses and so on. In 1512 in the rough draft of his first trea tise about proportions, Durer wrote: “All needs of the person are satiated by the fleeting things in the case of their surplus, so that they cause in him disgust, except for a thirst for knowledge.... A desire to know a lot and through this to grasp the essence of all things is inherent within us from nature.” These words became a prologue to Durer’s theoretical works. Art of that epoch Figure 3.30. Albrecht Durer’s is often penetrated by a thirst for knowledge and Melancholia Alexey Stakhov THE MATHEMATICS OF HARMONY 176 many representatives of that epoch became scientists and researchers. The idea of the unity of artistic inspiration and mathematical theory is also re flected in Durer’s wellknown picture Melancholia created in 1514. This pic ture embodies the image of a person, who is near to God. The individual is surrounded by various geometric tools (Fig. 3.30). The presence in the en graving of a polyhedron (most likely, truncated rhombohedra), is certainly not accidental. 3.8.4. Piero Della Francesca We should start a list of the greatest experts of the Renaissance, who made a concerted study of the geometry of polyhedra beginning with Piero Della Francesca (1420 1492). We know a little about the life of Piero Della Frances ca, who was a great Italian artist, art theorist and mathematician. It is known, that he was born into the family of a craftsman in the small city of Borgo San Sepulcro. He studied in Florence, then worked in some Italian cities, includ ing Rome. Francesca’s creativity went far beyond the local art schools and helped define the whole of art of the Italian Renaissance. Fortunately, at the beginning of the 20th century the originals of three mathematical manuscripts by Piero Della Frances ca were discovered. They are now housed in the Vatican Library. After five centuries of obscurity, the glory of this great mathematician has been returned to Piero Della Francesca. Now it is known for certain that Francesca was the Renaissance’s first expert, who rediscovered and described in detail all Archimedean Solids, in particular, the five trun cated Platonic Solids, the truncat ed tetrahedron, hexahedron, octa hedron, dodecahedron, and, what is especially important here, the trun cated icosahedron. Figure 3.31. Piero Della Francesca’s Piero Della Francesca was both The Baptism of Christ and its harmonious a talented mathematician and great analysis based on the golden section Chapter 3 Regular Polyhedrons 177 artist. His art works express majestic solemnity, nobleness and harmony of images, reasonableness of proportions, and clarity of perspective in their con struction. His painting of The Baptism of Christ is one of the highlights of his artistic style. Christ is placed in the center of the picture (Fig. 3.31). His feet are being washed by the river water; his hands are combined in a gesture of Catholic prayer. Near to Christ we see John the Baptist, who pours water from a dish on Christ’s head. The dove, representing the Holy Spirit, descends from above the head of Christ. 3.8.5. Art of Intarsia At the end of the 15th and beginning of the 16th centuries the art of Intar sia, a special kind of inlay, a mosaic, constructed from thousands of fine slices of various species of trees, was very pop ular in Northern Italy. The mosaic created by Fra Giovanni da Verona (1457 1525) for the church Santa Maria in Verona in roughly 1520, is an example of intarsia (Fig. 3.32). The image of halfopened shutters creates on the planar mosaic a spa tial effect, which increases through the representation of various poly hedra (including the truncated icosahedron) by using Leonardo’s Figure 3.32. Fra Giovanni da Verona’s intarsia technique of rigid edges. created for the church Santa Maria in Verona 3.8.6. Salvador Dali’s Last Supper Now, let us consider an example of the representation of polyhedra by the wellknown 20th century artist Salvador Dali (1904 1989). This great Spanish artist is one of the best known representatives of surrealism. An excellent art ist, Dali created images similar to dreadful visions, called by him “drawing pictures of dreams.” Some of the most common repeat images, for example hours, which lose their forms under the sun beams, became Dali’s logo. Dali’s creativity continues to cause disputes (e.g. some critics even suggest that af ter 1930 he did not create anything of real worth). However, in 1955 Dali created one of his best known pictures, Last Supper (Fig. 3.33). This big picture is an original masterpiece of painting. Geometrical ra tionalism testifies to the invincible belief in the sacred force of number. In Alexey Stakhov THE MATHEMATICS OF HARMONY 178 the center of the big horizontal picture (167×288) we see Christ. Together with his pupils he sits at the table as God’s son descend ed from the Heavens (Dodecahe dron) upon the Earth. The apos tles are represented with heads held low inclined to the table. They appeared as if in deep wor ship of Christ. The dodecahe Figure 3.33. Salvador Dali’s Last Supper dron in this picture plays a cen tral role because it personifies the spiritual harmony, moral cleanliness and greatness of character. 3.8.7. Escher’s Creative Work Maurits Cornelis Escher (1898 1972) is one of the world’s most famous graphic artists. His art continues to amaze millions of people all over the world. In his works we recognize his keen observation of the world around us and the expressions of his own fantasies. Escher shows us that reality is wonderful and fascinating. He is most famous for his socalled impossible structures, such as Ascending and Descending, Relativity, and his Transformation Prints such as Metamorphosis I, Metamorphosis II and Metamorphosis III, Sky & Water and Reptiles (see www.mcescher.com). Escher’s creative work was highly esteemed by many scientists, in partic ular, by mathematicians and crystallographers. At the International Crystal lographic Congress in Cambridge (1960) the exhibition of Escher’s pictures became a sensation attracting the special attention of the crystallographers. What do art and crystallography hold in common? The question arises as to whether Maurits Cornelis Escher intuitively discovered the laws of symme try, those laws which dominate over crystals and define their external shape, nuclear structure and physical properties, and then illustrated these laws in his pictures. Escher took great interest in periodic figures and drew up the mosaic patterns of the repeating figures in his pictures. He inserted one image in another so that identical figures are periodically repeated without leaving any empty space between them. In fact, it is based on the same law, according to which the particles in crystal structures are placed, namely, the law of the densest packing: the periodic recurrence of identical groups of particles with out intervals and infringements. Chapter 3 Regular Polyhedrons 179 3.9. Application of the Golden Mean in Contemporary Art 3.9.1 Abstract Art by Astrid Fitzgerald Astrid Fitzgerald is an internationally acclaimed artist, born and educated in Switzerland and now residing and working in New York. Her works are rep resented in major public museums and private collections in the United States, Europe and Asia. Fitzgerald’s installation, Amish Quilts, was chosen by the judges to represent the United States at the Artcanal Exposition (2002) in Switzerland. She is also the author of An Artist’s Book of Inspiration – A Collection of Thoughts on Art, Artists and Creativity (1996) and Being Consciousness Bliss – A Seeker’s Guide (2002), both published by Lindisfarne Books. Fitzgerald refers to her re cent work as Cosmic Measures – a phrase that expresses her continuing search for the true nature of things. Here she began to work with the fundamental laws of geo metric forms that include the Golden Mean – the uni versal principle that underlies Nature from the spiral of our DNA to that of our galaxy. These harmonious pro portions have fascinated philosophers, architects and artists for millennia. Fitzgerald embodies the golden mean within her abstract pictures. In Figs. 3.34 we can see some pictures of Astrid Astrid Fitzgerald Fitzgerald based on the Golden Mean. Figure 3.34. Abstract art of Astrid Fitzgerald Alexey Stakhov THE MATHEMATICS OF HARMONY 180 3.9.2. The World of Matjuska Teja Krasek Matjuska Teja Krasek obtained her B.A. degree in painting from the Arthose– College for visual arts in Ljubljana, Slovenia. Her theoretical and practical work focuses primarily on symmetry as a connecting concept between art and science. Through her art, Krasek wishes to convey her experience and feelings that are connected with her research results from various disciplines. She wants to convey this to all who are interested in exploring the way and functioning of our universe and nature. In this manner she wishes to contribute to the awareness of certain characteristics such as the various kinds of symmetry, the golden mean, and the Fibonacci sequence’s connection to nature, natural science and art. She also explores in her works how the use of various formal elements of artistic expression (lines, colors, structures, etc.) can influence the stability of art work. She uses contemporary computer technology as well as a classical painting technique. And her artworks have been represented at many international exhibitions and pub lished in international journals (Leonardo Journal, Leonardo online and so on). Matjuska Teja Krasek In Fig. 3.35 we can see some pictures of Matjuska Teja Krasek. Figure 3.35. The art works of Matjuska Teja Krasek 3.9.3. The Geometric Art of John Michell John Michell, born in 1933, was educated at Eton and Cambridge and published his first book in 1967. The author of 12 books, he is a specialist in Chapter 3 Regular Polyhedrons 181 sacred geometry and the geometry of reconciliation. With The View over Atlantis (1969) and City of Revelation (1972) Michell helped to change the worldviews of a whole gen eration by illuminating the science, culture and wisdom of past civilizations. Michell’s City of Revelation was sum marized and extended with further research in The Dimen sions of Paradise, in which the proportions and symbolic John Michell numbers of ancient cosmology were explained. Michell’s books, Twelve Tribe Nations and Science of Enchanting the Land scape (1991), the latter with Christine Rhone, explain that throughout the history of civilization an ideal social order harmonically related to nature and the zodiac was imposed upon the landscapes of the world. This is depicted, for example, by the 12 tribe divisions of people and the careful alignment of var ious holy places. Fig. 3.36. The Geometric Art of John Michell 3.9.4. Quantum Connections by Marion Drennen Marion Drennen is a Louisiana Artist, who received her Bachelor in Fine Arts from LSU, working in Acrylic on Board. These paintings incorporate the concepts of Number and Quantum Physics in an effort to evoke a sense of con nectedness across time and space, within ourselves and in our relationships. In her artistic statement she describes her artistic process – “Being an idea person, an avid reader and re searcher, my paintings manifest from contemplations on a variety of subjects – Mathematics, The Golden Ratio, Quantum Physics, and Spiritual. I glimpse the connec tions and begin to draw thumbnail sketches around the edges of my notes. The internal dialogue is about the con Marion Drennen cept, then words gradually disappear and visual elements Alexey Stakhov THE MATHEMATICS OF HARMONY 182 expand until only sketches are coming off the end of the pencil. Later, within the painting, words may reappear.” She compares her creation process with a dance – “I used to dance. There’s something exquisite about losing yourself in movement. Sometimes, when I’m painting, I’ll start to dance. I almost always paint to music. After working with the Golden Ratio for several years, I now have custom painting surfaces made that allow me to paint within that for mat. It sets the stage. It is the architecture, the structure on which I begin to build. I break up the space and then insert my one or two shapes, the initial idea, and then the dance begins.” In Fig. 3.37 we represent some pictures of Marion Drennen from her Quan tum Connections exhibit that was showing at the Brunner Gallery in the Shaw Center for the Arts. Fig. 3.37. Quantum Connections by Marion Drennen. 3.10. Conclusion The regular and semiregular polyhedra have been known from antiquity. The regular polyhedra got the name Platonic Solids, because they played such an important role in Plato’s cosmology. According to Plato, the atoms of the Universe’s “Basic Elements” have the form of the Platonic Solids (Fire – a Tet rahedron, Earth – a Hexahedron or Cube, Air – an Octahedron, Water –an Icosahedron). The Dodecahedron was considered to be the primary figure of the Universe, expressing the Universal Intellect and the Harmony of the Uni verse. The semiregular polyhedra are named Archimedean Solids. A geometric theory of the Platonic Solids was presented in the 13th or final Book of Euclid’s Chapter 3 Regular Polyhedrons 183 Elements. This is the reason why the ancient Greek mathematician Proclus, a commentator on Euclid, put forward the hypothesis about the true purpose of Euclid writing The Elements. In Proclus’ opinion, Euclid wrote his Elements to give a full and systematized theory of geometric construction of the “ideal” geo metric figures, in particular, the five Platonic Solids. Thus, we can consider The Elements of Euclid to be the first historical geometrical theory of the Harmony of the Universe, based upon the golden section (division in the extreme and mean ratio) and the Platonic Solids! Since antiquity the Platonic and Archimedean Solids have been a source of many scientific hypotheses, theories and discoveries. The surprising coincidence of the numerical characteristics of the dodecahedron (12 faces, 30 edges and 60 planar angles on its surface) with the main cycles of the Solar System (12year cycle of Jupiter, 30year cycle of Saturn, and 60year basic cycle of the Solar System) apparently became one reason why the numbers 12, 30, 60 and 360=12×30 were used by the Ancient Egyptians in their calendar and their systems of time and angle measurement. In the 19th century the prominent mathematician Felix Klein began to consid er the regular Icosahedron as the main geometric object, from which the branches of the five mathematical theories follow, namely, geometry, Galois’ theory, group theory, theory of invariants and differential equations. Shechtman’s quasicrys tals were based on the icosahedron and fullerenes (1996 Nobel Prize in chemis try). These are brilliant confirmations of the role of the Platonic and Archimedean Solids, and therefore the golden section, in modern physics. This page intentionally left blank Part II Part II. Mathematics of Harmony Alexey Stakhov THE MATHEMATICS OF HARMONY 186 Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 4.1. A Combinatorial Approach to the Harmony of Mathematics 4.1.1. Mathematical, Aesthetic and Artistic Understanding of Harmony In the Introduction we mentioned that the Harmony Problem is one of the “key” problems of mathematics entering the scene at its very origin. But what does a concept of “harmony” mean? The Russian philosopher Shestakov, one of the best researchers in the field, pointed out three basic understandings of “harmony” that have been devel oped in science and aesthetics since antiquity [7]: 1. Mathematical Understanding of Harmony or Mathematical Harmony. In this sense, harmony is understood as equality or proportion ality of parts one to another and the parts to the whole. 2. Aesthetic Harmony. In contrast to the mathematical harmony, the aesthetic harmony is not quantitative, but qualitative notion and express es the internal nature of things. The aesthetic harmony is connected with aesthetic excitements and estimations. Most precisely this type of harmo ny is shown at perception of beauty of Nature. 3. Artistic Harmony. This type of harmony is connected with art. Artistic harmony is an actualization of the harmony principle in the realm of art. In the present book, our attention is concentrated on Mathematical Harmony. It is clear that mathematical harmony is expressed in the form of certain numerical proportions. Shestakov emphasizes that mathematical har mony “attracts attention to its quantitative side and is indifferent to qualita tive originality of the parts forming conformity.... The mathematical under standing of the harmony fixes, first of all, quantitative definiteness of the har mony, but it does not express aesthetic quality of the harmony, its expressive connection with beauty.” Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 187 Thus, in this book we consciously restrict the area of our research to the mathematical understanding of harmony. This approach is fruitful in mathe matics and allows us to create a Mathematical Theory of Harmony or Mathe matics of Harmony that extends the area of the mathematical models of the harmonic processes of Nature. 4.1.2. Concept of the Mathematical Theory of Harmony We can ask a question: how does one create a Mathematical Theory of Harmony? As is known, mathematics studies the quantitative aspect of this or that phenomenon. Starting with a mathematical analysis of the concept of harmony, we should concentrate our attention on the quantitative aspects of harmony. What is the quantitative aspect of this concept? There are many definitions of the harmony concept. However, the major ity of them are reduced to the following definition that is given in The Great Soviet Encyclopedia: “Harmony is proportionality of parts and the whole, a combination of the various components of the object in the uniform organic whole. It is the inter nal order and measure obtained in the harmony external expression [of the harmony].” In the article Harmony in Nature and Art [149], the Russian crystallogra pher Shubnikov compared harmony with the Order studied by science to dis cover Nature’s laws. He writes: “Law, Harmony, Order underlie not only sci entific work but also any work of art.” Now let us analyze a question of the origin and meaning of the word “harmo ny.” “Harmony” has a Greek origin. The Greek word αρµουια has the following meaning: connection, consent. The analysis of the word of “harmony” and its defi nitions demonstrate that the most important, key notions, which underlie this concept, are the following: Connection, Consent, Combination, Order. We can ask a question: what branch of mathematics studies these concepts? A search for the answer leads us to Combinatorial Analysis. “Combinatorial anal ysis studies the various kinds of combinations and connections, which can be formed from the elements of some finite set. The term ‘combinatorial’ is derived from the Latin word combinary which means to combine or connect.” [150] It follows from this consideration that both the Latin word combinary and the Greek word αρµουια, in essence, have the same meaning, namely, Combi nation and Connection. It allows us to put forward the following hypothesis: the “Laws of combinatorial analysis” can be used for the analysis of the Har mony concept from the quantitative point of view. Alexey Stakhov THE MATHEMATICS OF HARMONY 188 Therefore, the main concepts underlying the Harmony Mathematics and the combinatorial analysis, in essence, coincide. And this allows us to use com binatorial analysis as the main mathematical theory for the analysis of the harmony concept from the quantitative point of view. This is the main idea behind the research, in which the author made an attempt to create the Fun damentals of the Harmony Mathematics based upon combinatorial analysis. Thus, Harmony Mathematics is a mathematical theory studying the no tion of harmony from a quantitative point of view. Its main goal is to study mathematical laws, mathematical proportions, which underlie the Harmony of the Universe. 4.1.3. An Analogy between the Theory of Information and Mathematics of Harmony Are there similar theories in modern science? The Mathematical Theory of Information developed by Claude Shannon [151] is possibly a brilliant exam ple of a similar theory. Information is a complex and interdisciplinary concept similar to Harmony. Both information and harmony are nonmaterial and om nipresent. In spite of its nonmaterial character, the Mathematical Theory of Harmony may be considered to be an original mathematical theory similar to the Theory of Information. By developing a Theory of Information, Claude Shannon used the concept of Probability as the starting point of the theory. The concept of Entropy, based on the notion of probability, became the underlying basic concept of the The ory of Information. It is necessary to emphasize that Shannon’s Theory of In formation is a mathematical theory and can be effectively used, first of all, for the quantitative analysis of any informational system. In his wellknown arti cle Bandwagon [151] Shannon warned that one must be cautious applying this theory to other areas of human activity. Shannon’s Theory of Information is sometimes considered to be a branch of probability theory. By continuing the analogy between Shannon’s Theory of Information [151] and Harmony Mathematics, it is possible to consider the Harmony Mathematics as a special branch of combinatorial analysis. From this point of view, it becomes imperative to answer the question about practical application of the Harmony Mathematics. What are the areas of the effective application of this theory? By answering this question, it is important to emphasize that the most effective areas are those where the quan titative aspects of the Harmony are most important, such as theoretical phys ics, computer science, biology, botany, economics, and so on. Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 189 4.2. Binomial Coefficients and Pascal Triangle 4.2.1. The Main Concepts of Combinatorial Analysis Combinatorial analysis studies different kinds of combinations that can be built up from the elements of some set. The term “combinatorial analysis” originates from the Latin word combinare that has the meaning to combine or connect. Some elements of combinatorial analysis were known in India already in the 2nd century AD. The Indian mathematicians knew how to calculate num bers Cnm called Combinations from n elements by m, and they knew the fol lowing formula: (4.1) Cn0 + Cn1 + ... + Cnn = 2n . A theory of binary codes, the basis of modern computers, is an example of the effective application of the formula (4.1). We can consider a set of the ndigit binary words starting from the code combination 00 … 0 and ending by the code combination 11 … 1. As is known, the number of elements of this set is equal to 2n. We can divide this set into the (n+1) disjoint subsets. Then we can refer all binary words consisting only of 0’s to the first subset. It is clear that the only code combination 00 … 0 satisfies this condition, that is, the number of the elements of this subset is equal to Cn0 = 1. Then, we can refer to the second subset all the binary words containing only 1 and the (n1) 0’s. It is clear that the number of elements of this subset is equal to Cn1 . We can refer to the (m+1)th subset all the ndigit binary words containing m 1’s and (n m) 0’s; the number of code combinations of this subset is equal to Cnm . At last, we can refer to the (n+1)th subset all the ndigit binary words which contain only 1’s. It is clear that the only code combination 11 … 1 satisfies this condi tion, that is, the number of elements of this subset is equal to Cnn = 1. From this reasoning the validity of the formula (4.1) follows. The term “combinatorial analysis” began to be used after publication in 1666 of Leibniz’s Reasoning about Combinatorial Art; in this book he gave for the first time scientific substantiation of the theory of combina tions and permutations. Bernoulli introduced for the first time a notion of Distribution in the second part of his famous book Art of Guessing pub lished in 1713. He introduced and used in our sense the term Permuta tion. The term Combination was introduced by Pascal in his Treatise about the Arithmetical Triangle (1665). Alexey Stakhov THE MATHEMATICS OF HARMONY 190 Thus we have the wellknown formula Cnm = n! m !(n − m)! , (4.2) where n!=1×2×3×…×n is a Factorial of n. 4.2.2. Binomial Formula Now, write the following formulas: (a + b) = 1 1 ( a + b ) = 1a + 1b ( a + b )2 = 1a2 + 2ab + 1b2 3 ( a + b ) = 1a3 + 3a2b + 3ab2 + 1b3 . 0 (4.3) Note that the first two formulas from (4.3) are trivial; the other two for mulas are wellknown from secondary school. We can ask the question: how do we calculate the binom (a+b)n? The wellknown mathematical formula called the Binomial Formula gives the an swer to this question: ( a + b ) = a n + Cn1 a n −1b + Cn2 a n −2b2 + ... + Cnk a n − k b k + ... + Cnn −1abn −1 + bn . n Cnk (4.4) are named Binomial Coefficients or Binomial Here the numbers Factors. Sometimes the discovery of formula (4.4) was attributed to Newton. How ever, long before Newton, the mathematicians of many countries, in particu lar, the Arab mathematician Al Kashi, the Italian mathematician Tartalja, the French mathematicians Fermat and Pascal, knew this formula. Newton’s mer it consists in the fact that he derived this formula for the case of any real num ber n, that is, he proved that the formula (4.4) is true when n is rational or irrational, positive or negative. The formulas (4.3) are special cases of a general formula (4.4). In particu lar, for the case n=1 the formula (4.4) is reduced to the following: ( a + b )1 = C10 a + C11b = 1a + 1b, whence it appears that C10 = 1 and C11 = 1. For the case n=2, the formula (4.4) takes the following form: ( a + b ) = C20 a2 + C21ab + C22 = 1a2 + 2ab + 1b2 , 2 whence it appears that C20 = 1, C21 = 2, C22 = 1. Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 191 Thus, we can get the decomposition (4.4) which can be very easy if we know how to calculate the binomial coefficients Cnk . 4.2.3. Pascal Triangle The French 17th century mathematician Blaise Pascal (16231662) sug gested an original method for the calculation of the coefficients Cnk for any arbitrary nonnegative integers n and k. In addition to the properties (4.1) and (4.2), the binomial coefficients Cnk have a number of remarkable properties that are given here without a proof: Cn0 = Cnn = 1 (4.5) Cnk = Cnn − k (4.6) (4.7) Cnk+1 = Cnn −1 + Cnk . The last property (4.7) is also named the Pascal Law. Using the recursive relation (4.7), Pascal had offered an orig inal method for the calculation of binomial factors that are based on their disposition in the form of a special numerical table called Pascal’s Triangle. Let us examine an infinite table of numbers constructed according to the Pascal Law (4.7). Blaise Pascal (16231662) The top of the indicated table (Fig. 4.1) that is named Zerorow, consists of the only binomial coefficient C00 = 1. The next row – the 1st row consists of two binomial coefficients C10 = C11 = 1. Each succeeding row can be constructed from the preceding row according to the rules (4.5) (4.7). It is easy to prove the following properties of Pascal triangle: 1. The sum of the binomial coefficients of the nth row of Pascal triangle is equal to 2n what corresponds to the identity (4.1). 2. All rows of Pascal triangle are symmetric relative to the binomial coef ficient C00 = 1 of the zerorow that corresponds to the property (4.6). The above Pascal Trian gle appeared for the first time 1 1 1 in Pascal’s Treatise about Ar 1 2 1 ithmetical Triangles written in 1 3 3 1 1 4 6 4 1 1665. However, one century 1 5 10 10 5 1 prior to the publication of 1 6 15 20 15 6 1 1 7 21 35 35 21 7 1 Pascal’s Treatise, this numer 1 8 28 56 70 56 28 8 1 1 9 36 84 126 126 84 36 9 1 ical table (but in rectangular Figure 4.1. Pascal Triangle form rather than triangular) Alexey Stakhov THE MATHEMATICS OF HARMONY 192 was described in the General Treatise about Number and Measure written by the Italian mathematician Nikola Tartalja (15001557). Tartalja’s Treatise was published immediately after his death. Here the first or top row Table 4.1. Tartalja’s Rectangle and the first column on the 1 1 1 1 1 1 left consist only of 1’s; each 1 2 3 4 5 6 “internal” number of other 1 3 6 10 15 21 1 4 10 20 35 56 rows is equal to the sum of 1 5 15 35 70 126 two numbers, the first one 1 6 21 56 126 252 that stands in the same row on the left to it and the second one that stands in the preceding row above it. This table of binomial factors is called Tartalja’s Rectangle (Table 4.1). 4.3. The Generalized Fibonacci pNumbers 4.3.1. Rectangular Pascal Triangle There are many forms of the representation of the Pascal Triangle, for exam ple, in the form of an isosceles triangle, in the form of a rectangular table (Tartal ja’s Rectangle), etc. We will examine the socalled Rectangular Pascal Triangle that can be represented by the following table of binomial coefficients (Fig. 4.2). The rows of the Pascal Triangle are numbered C 0 C 0 C 0 C 0 C 0 C 0 … C 0 n 0 1 2 3 4 5 1 1 1 1 1 1 from top to bottom. The binomial coefficients C1 C 2 C 3 C 4 C 5 … C n 2 2 2 2 2 C00 = C10 = C20 = ... = Cn0 = 1 make up a “zero” row. C2 C3 C4 C5 … C n 3 3 3 3 Every nth row starts with the binomial coefficient C3 C4 C5 … C n 4 4 4 of the kind Cnn = 1 ( n = 0, 1, 2, 3, ...) . C4 C5 … C n 5 5 The columns of the Pascal Triangle are num C5 … C n Figure 4.2. bered from left to right; the first lefthand column Rectangular Pascal % # n Cn Triangle that consists of the only binomial coefficient C00 = 1 is called the Zerocolumn. The nth col umn (n=0,1,2,3,…) includes the following binomial coefficients: ( ) Cn0 , Cn1 , Cn2 , ..., Cnk , ..., Cnn − k , ..., Cnn , where Cnk = Cnn − k . As mentioned above, Pascal Triangle is based on the recursive relation (4.7). What is the correlation between Pascal Triangle and Fibonacci numbers? In the second half of the 20th century many Great mathematicians (Martin Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 193 Gardner [12], George Polya [17], Alred Renyi [23] and others) independently one after another discovered the connection of Fibonacci numbers with Pascal’s Triangle and binomial coefficients. This discovery demonstrates a fundamental connection of the Harmony Mathematics based on Fibonacci numbers and the golden mean with combinatorial analysis and outlines a way for the future gen eralization of Fibonacci numbers. Below we will demonstrate a surprisingly sim ple mathematical regularity that connects Pascal Triangle and Fibonacci num bers. A generalization of this regularity resulted in the mathematical discovery called the Generalized Fibonacci pNumbers [19, 20]. 4.3.2. Pascal pTriangles and Fibonacci pNumbers Let us examine the Rectangular Pascal Triangle represented in numerical form (Fig. 4.3). We can name the given table of binomial co 1 1 1 1 1 1 1 1 1 1 efficients Pascal 0Triangle (the meaning of this 1 2 3 4 5 6 7 8 9 1 3 6 10 15 21 28 36 definition will become clear below). If we sum the 1 4 10 20 35 56 84 binomial coefficients of the Pascal 0Triangle by 1 5 15 35 70 126 columns starting from the 0column, 1 6 21 1 56 126 1 7 28 84 then according to (4.1) we obtain the binary se 1 8 36 quence: 1 9 1 1 2 4 8 16 32 64 128 256 512 (4.8) 1, 2, 4, 8, 16, ..., 2n , .... Now, we do some “manipulations” around the Figure 4.3. Pascal 0Triangle Pascal 0Triangle. We move each row of the Pas cal 0Triangle one column to the right with respect to the previous row. As a result of such move, we obtain a table called Pascal 1Triangle (Fig. 4.4). Now, sum the binomial coefficients of the Pascal 1Triangle in each column. To our amazement, we find that this summation results in the Fibonacci numbers: 1,1,2,3,5,8,13,21,…,Fn+1,…. (4.9) where Fn+1 is the (n+1)th Fibonacci number given by the following recursive relation: (4.10) Fn+1= Fn+Fn1. The numerical sequence (4.9) is generated by the recursive relation (4.10) at the seeds: F1=F2=1. (4.11) If we move in the initial Pascal 0Triangle (Fig. 4.1) the binomial coeffi cients of each row by p columns to the right with respect to the previous row Alexey Stakhov THE MATHEMATICS OF HARMONY 194 (p=0,1,2,3,…), we get the numerical table named 1 1 1 1 1 1 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 10 Pascal pTriangle. It is clear that the Pascal 0 1 3 6 10 15 21 28 36 Triangle that corresponds to the case p=0 is the 1 4 10 20 35 56 initial Pascal Triangle (Fig. 4.3). The Pascal 1 1 5 15 35 1 6 Triangle is represented in Fig. 4.4. The Pascal 1 1 2 3 5 8 13 21 34 55 89 144 pTriangles for the cases p=2 and p=3 have the forms shown in Fig. 4.5. Figure 4.4. Pascal 1Triangle Now, sum the binomial coefficients in each column of the Pascal 2 and 3 Triangles. As a result, we obtain two new numerical sequences: 1, 1, 1, 2, 3, 4, 6, 9, 13, 19, 28, 41, 60, ... (4.12) 1, 1, 1, 1, 2, 3, 4, 5, 7, 10, 14, 19, 26, ... (4.13) Pascal 2Triangle Pascal 3Triangle 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 10 1 3 6 10 15 21 28 1 4 10 20 1 1 1 1 2 3 4 6 9 13 19 28 41 60 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 1 3 6 10 15 1 1 1 1 1 2 3 4 5 7 10 14 19 26 Figure 4.5. Pascal 2Triangle and 3Tirangle Denote by F2(n) and F3(n) the nth elements of the sequences (4.12) and (4.13), respectively. It is easy to see the following regularities in the numerical sequences (4.12) and (4.13) that can be expressed by the following recursive relations F2(n)= F2(n1)+F2(n3) for n≥4 at the seeds (4.14) F2(1)=F2(2)=F2(3)=1 and by the recursive relation (4.15) F3(n)=F3(n1)+F3(n4) for n≥5 at the seeds (4.16) (4.17) F3(1)=F3(2)=F3(3)=F3(4)=1. Thus, as a result of this examination we have found two new numerical se quences. The first of them that is given by the recursive relation (4.14) at the seeds (4.15) is named the Fibonacci 2numbers and the second one that is given by (4.16) at the seeds (4.17) is named the Fibonacci 3numbers. In the general case, for arbitrary р if we sum the binomial coefficients of each column of the Pascal pTriangle, we obtain the numerical sequences giv en by the following recursive relation: Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 195 Fp(n)=Fp(n1)+Fp(np1) for n≥p+1 at the seeds (4.18) (4.19) Fp(1)=Fp(2)=...=Fp(p+1)=1. The numerical sequences that correspond to the recursive relation (4.18) and (4.19) are named [20] the Fibonacci pnumbers. 4.3.3. Partial Cases of the Fibonacci pNumbers It is clear that for the case p=0 the recursive relation (4.18) and the seeds (4.19) takes the following form: F0(n)=F0(n1)+F0(n1) for n ≥2 (4.20) F0(1)=1. (4.21) It is easy to guess that the recursive relation (4.20) at the seed (4.21) generates the binary sequence (4.8) that is a special case of the Fibonacci pnumbers for p=0. Let us examine the case p=1. For this case the recursive relation (4.18) and the seeds (4.19) are reduced to the following: F1(n)=F1(n1)+F1(n2) for n≥3 (4.22) F1(1)=F1(2)=1. (4.23) Comparing these formulas with the recursive relation for the classical Fi bonacci numbers (4.10) and (4.11), we can conclude that the Fibonacci 1 numbers coincide with the classical Fibonacci numbers, that is, F1(n)=Fn. At last, we find that for the case p=∞ the sequence of the Fibonacci p numbers consists only of the unities: {1,1,1,…}. 4.3.4. A Representation of the Fibonacci pNumbers by the Binomial Coefficients In the above we have examined the formula (4.1) that allows us to repre sent binary numbers by the binomial coefficients. Analyzing the Pascal 1Tri angle (Fig. 4.4), it is easy to derive the mathematical formula that allows us to represent the Fibonacci 1numbers by the binomial coefficients: F1 ( n + 1) = Cn0 + Cn1 −1 + Cn2−2 + Cn3−3 + Cn4−4 + .... (4.24) This means that there are two ways to calculate the Fibonacci numbers, namely, by using the recursive relation (4.10) at the seeds (4.11) or by using the formula (4.24). Alexey Stakhov THE MATHEMATICS OF HARMONY 196 For example, using the formula (4.24), we can represent the Fibonacci num ber F1(7)=13 as follows: (4.25) F1 ( 7 ) = C60 + C51 + C42 + C33 + C24 + .... 4 Note that the binomial coefficient C2 in the sum (4.25) as well as all of the binomial coefficients following C24 are equal to 0 identically. This means that the expression (4.25) is the following finite sum of the binomial coefficients: (4.26) F1 ( 7 ) = C60 + C51 + C42 + C33 = 1 + 5 + 6 + 1 = 13. Studying the Pascal pTriangle, we can represent the Fibonacci pnum ber Fp(n+1) by the binomial coefficients as follows: Fp ( n + 1) = Cn0 + Cn1 − p + Cn2−2 p + Cn3− 3 p + Cn4− 4 p + .... (4.27) Note that the known formula (4.1) is a partial case of (4.27) for p=0 and the formula (4.24) is a partial case of (4.27) for p=1. 4.3.5. The “Extended” Fibonacci рNumbers Until now we have studied the Fibonacci pnumbers Fp(n) given for the positive values of n. In Chapter 2 we extended the classical Fibonacci num bers into the side of the negative values of n. By analogy, we can find the ex tended Fibonacci pnumbers, if we extend the Fibonacci pnumbers to the side of the negative values of n. With this purpose, we will find some general properties of such extended sequences. For the calculation of the Fibonacci p numbers Fp(n) corresponding to the nonnegative values of n=0,1,2,3,… we use the recursive relation (4.18) and the seeds (4.19). Let us represent the Fibonacci pnumber Fp(p+1) in the form (4.18) as follows: Fp(p+1)=Fp(p)+Fp(0). (4.28) According to (4.19), we have: Fp(p+1)=Fp(p)=1. This means that Fp(0)=0. Continuing this process, that is, representing the Fibonacci pnumbers Fp(p), Fp(p1),…, Fp(2) in the form (4.18), we have: Fp(0)=Fp(1)=Fp(2)=...=Fp(p+1)=0. (4.29) Now, represent the Fibonacci pnumber Fp(1) in the form (4.18): Fp(1)=Fp(0)+Fp(p). As Fp(1)=1 and Fp(0)=0, we get from (4.30): (4.30) (4.31) Fp(p)=1. Representing the Fibonacci pnumbers Fp(0), Fp(1), …, Fp(p+1) in the form of (4.18) , we get: Fp(p1)=Fp(p2)=...=Fp(2p+1)=0. (4.32) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 197 Table 4.2. The “Extended” Fibonacci рNumbers N 6 5 4 3 2 1 0 1 2 3 4 5 6 7 F1(n) 8 5 3 2 1 1 0 1 1 2 3 5 8 13 F2(n) 4 3 2 1 1 1 0 0 1 0 1 1 1 2 F3(n) 3 2 1 1 1 1 0 0 0 1 0 0 1 1 F4(n) 2 1 1 1 1 1 0 0 0 0 1 0 0 0 F5(n) 1 1 1 1 1 1 0 0 0 0 0 1 0 0 By continuing this process, we can obtain all values of the Fibonacci pnumbers Fp(n) for the negative values of n. Table 4.2 gives some values of the “ex tended” Fibonacci pnumbers for the cases p=1, 2, 3, 4, 5. 4.3.6. Some Identities for the Sums of the Fibonacci рNumbers Once again, let us examine the recursive relation (4.14). Decomposing the Fibonacci 2number F2(n1) in (4.14) according to the same recursive re lation (4.14), that is, representing F2(n1) in the form F2(n1)=F2(n2)+F2(n 4), we can represent the recursive relation (4.14) as follows: F2(n)=F2(n2)+F2(n3)+F2(n4). (4.33) This means that the sum of the three successive Fibonacci 2numbers is always equal to the Fibonacci 2number, which is two positions from the se nior Fibonacci 2number of the sum. Now, consider the recursive relation for the Fibonacci 3numbers given by (4.16). Decomposing the Fibonacci 3number F3(n1) in (4.16) according to the same recursive relation (4.16), we can represent the recursive relation (4.16) as follows: (4.34) F3(n)=F3(n2)+F3(n4)+F3(n5). Decomposing the number F3(n2) in (4.34) according to the recursive re lation (4.16), we can represent (4.34) as follows: F3(n)=F3(n3)+F3(n4)+F3(n5)+F3(n6), (4.35) that is, the sum of the four sequential Fibonacci 3numbers is always equal to the Fibonacci 3number, which is three positions from the senior Fibonacci 3 number of the sum. If we use a similar approach to the Fibonacci рnumbers in the general case, we obtain the following general identity: (4.36) Fp(n)=Fp(np)+Fp(np1)+Fp(np2)+...+Fp(n2p). Note that the identity (4.36) is valid for all “extended” Fibonacci pnum bers Fp(n) when n takes the values from the set: {0,±1,±2,±3,…}. Now, let us consider the sum of the first n Fibonacci pnumbers: Fp(1)+Fp(2)+Fp(3)+...+Fp(n). (4.37) Alexey Stakhov THE MATHEMATICS OF HARMONY 198 To get the required result, we write the basic recursive relation (4.18) for the Fibonacci pnumbers as follows: Fp(n)=Fp(n+p+1)Fp(n+p). By using (4.38), we can write the following equalities: (4.38) Fp (1) = Fp ( 2 + p ) − Fp (1 + p ) Fp ( 2 ) = Fp ( 3 + p ) − Fp ( 2 + p ) Fp ( 3 ) = Fp ( 4 + p ) − Fp ( 3 + p ) ... Fp ( n − 1) = Fp ( n + p ) − Fp ( n + p − 1) Fp ( n ) = Fp ( n + p + 1) − Fp ( n + p ) . Summing term by term the lefthand and righthand parts of these equal ities and taking into consideration that Fp(1+p)=1, we obtain the following expression for the sum (4.37): (4.39) Fp(1)+Fp(2)+Fp(3)+...+Fp(n)=Fp(n+p+1)1. The formula (4.39) includes a further number of remarkable formulas of discrete mathematics. In fact, for the case p=0 this formula is reduced to the following wellknown formula for binary numbers: 20+21+22+…+2n1=2n1. For the case p=1, the Fibonacci рnumbers coincide with the classical Fi bonacci numbers, that is, F1(n)=Fn. And then the formula (4.39) is reduced to the following formula: F1+F2+F3+...+Fn=Fn+21, that is well known from Fibonacci number theory [13, 16]. Thus, our “manipulations” with Pascal’s Triangle resulted in a small math ematical discovery! We found an infinite number of the new numerical se quences named the Fibonacci pnumbers (p=0,1,2,3,…). These numerical se quences include the binary numbers (4.8) (p=0) and the classical Fibonacci numbers (4.9) (p=1) as partial cases. These numerical sequences possess a number of interesting mathematical properties, and their study can result in widening the Fibonacci number theory. 4.3.7. The Ratio of Adjacent Fibonacci pNumbers In Chapter 2 we found that the classical Fibonacci numbers are closely connected with the golden mean. In particular, the limit of the ratio Fn /Fn1 aims for the golden mean. There is a question: what is a limit of the ratio of the two adjacent Fibonacci pnumbers? Introduce the following definition: Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 199 Fp (n) lim n→∞ F (n − 1) p = x. (4.40) By using the recursive relation (4.18), we can represent the ratio of the two adjacent Fibonacci pnumbers as follows: Fp (n) Fp (n − 1) =1+ = Fp (n − 1) + Fp (n − p − 1) Fp (n − 1) 1 Fp (n − 1) Fp (n − p − 1) 1 Fp (n − 1) ⋅ Fp (n − 2) ⋅⋅⋅ Fp (n − p) =1+ . (4.41) Fp (n − 2) ⋅ Fp (n − 3) " Fp (n − p − 1) Taking into consideration the definition (4.40), for the case n→∞ we can replace the expression (4.41) by the following algebraic equation: xp+1=xp+1. (4.42) Note that the equation (4.42) is called the Characteristic Equation for the Recursive Relation (4.18). Denote by τp a positive root of the characteristic equation (4.42). Let us examine Eq. (4.42) for the different values of p. For p=0, Eq. (4.42) is reduced to the trivial case: x=2. For p=1, Eq. (4.42) is reduced to the classical golden algebraic equation: x2=x+1 (4.43) with a positive root τ = 1 + 5 2. Thus, Eq. (4.42) can be considered to be a very broad generalization of the golden equation (4.43). ( ) 4.4. The Generalized Golden pSections 4.4.1. A Generalization of the Division in Extreme and Mean Ratio (DEMR) Equation (4.42) has the following geometric interpretation. Let us give the integer р a nonnegative value (p=0,1,2,3,…) and divide the line АВ at the point С in the following proportion (Fig. 4.6): p CB AB . = AC CB (4.44) Alexey Stakhov THE MATHEMATICS OF HARMONY 200 C a) p = 0 A C b) p = 1 A e) p = 4 A B τ1=1.618 C c) p = 2 A d) p = 3 A B τ0=2 B τ2=1.465 C B τ3=1.380 C B τ4=1.324 Figure 4.6. The Generalized Golden pSections Note that the proportion (4.44) is reduced to the “dichotomy” for the case p=0 (Fig. 4.6a) and to the classical “division in the extreme and mean ratio” (the golden section) for the case p=1 (Fig. 4.6b). Taking this fact into consid eration, we will name the division of the line segment AB at the point C in the proportion (4.44) a Golden pSection and the positive root of Eq. (4.42) a Gold en pProportion [20]. 4.4.2. Algebraic Properties of the Golden рProportions If we substitute the golden рproportion τp for x in Eq. (4.42), we get the following identity for the golden рproportion: τ pp +1 = τ pp + 1. (4.45) If we divide all terms of the identity (4.45) by τ pp , we get the following remarkable property of the golden рproportion: 1 τp = 1+ p τp (4.46) or τp −1 = 1 . τ pp (4.47) Note that for the case p=0 (τp=2) the identities (4.46) and (4.47) are reduced to the following trivial cases: 1 1 2 = 1 + or 2 − 1 = . 1 1 ( ) For the case p=1 we have τ1 = τ = 1 + 5 2 and the identities (4.46), (4.47) are reduced to the identities (1.11) and (1.12). Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 201 If we multiply and divide repeatedly all terms of the identity (4.45) by τp, we obtain the following remarkable identity connecting the powers of the gold en pproportion: τnp = τnp−1 + τnp− p −1 = τ p × τnp−1 . (4.48) Note that for the case p=0 the identity (4.48) is reduced to the following trivial identity for the binary numbers: 2n=2n1+2n1=2×2n1. For the case p=1 the identity (4.48) is reduced to the following wellknown identity for the classical golden mean: τn=τn1+τn2=τ×τn1. (4.49) 4.4.3. Geometric Progressions Based on the Golden pProportions Now, consider a geometric progression based on the golden pproportion: {τ , τ , ..., τ , τ , τ , τ = 1, τ , τ , τ , ...}. n p n −1 p 3 p 2 p 1 p −1 p 0 p −2 p −3 p (4.50) The geometric progression (4.50) possesses the remarkable property: for the case p>0 according to (4.48) each term of the geometric progression (4.50), for example, τnp can be obtained from the preceding terms in two ways: (1) by the multiplication of the preceding term by τp τnp = τ p × τnp−1 ; (2) by the summation of the (n1)th and the (np1)th terms τnp = τnp−1 + τnp− p −1 . Note that till now we believed that only the golden geometric progression based on the classical golden mean possesses similar “additive” property. It follows from this consideration that a number of similar geometric progres sions are infinite and all of them are based on the golden pproportions. By decomposing τnp and all the cases arising from such decomposition terms τnp−1 , τnp−2 , τnp− 3 , ... according to the recursive relation (4.48), we obtain the fol lowing identities: τnp = τnp− p −1 + τnp−1 ( ( ) ) τnp = τnp− p −1 + τnp− p −2 + τnp−2 τnp = τnp− p −1 + τnp− p −2 + τnp− p −3 + τ −p3 ... (4.51) k n n− p− j n−k τp = τp + τp . j =1 In particular, for the case k=p the identity (4.51) takes the following form: p p τnp = τnp− p − j + τnp− p = τnp− p − j . (4.52) j =1 = 0 j ∑ ∑ ∑ Alexey Stakhov THE MATHEMATICS OF HARMONY 202 By decomposing τnp and terms τnp−( p +1) , τnp−2( p +1) , τnp−( k −1)( p +1) ,... arising from such decomposition according to the recursive relation (4.48), we obtain the following identities: τnp = τnp−1 + τnp−( p +1) τnp = τnp−1 + τnp−( p +1)−1 + τnp−2( p +1) τnp = τnp−1 + τnp−( p +1)−1 + τnp−2( p+1)−1 + τ np−3( p+1) ... k τ np = τ np−( j −1)( p+1)−1 + τnp− k( p +1) . j =1 (4.53) ∑ In particular, for the case p=0 (τp=2) the identities (4.51) and (4.53) co incide and they are reduced to the following remarkable identity for the “bi nary” numbers: k 2n = ∑2 n− j + 2n − k . j =1 (4.54) ( ) For the case p=1 we have: τ1 = τ = 1 + 5 2 ; then the identities (4.51) and (4.53) take the following forms, respectively: τn = τn = k ∑τ n − j −1 ∑τ n − 2 j +1 j =1 k + τn − k + τn −2 k . j =1 (4.55) (4.56) 4.5. The Generalized Principle of the Golden Section 4.5.1. Dichotomy Principle The remarkable book [46] by Russian architect Shevelev is devoted to a study of the most general principles that underlie Nature. The Dichotomy Prin ciple and the Golden Section Principle are the most important of them. The Dichotomy Principle is based on the following trivial property of the binary numbers: 2n=2n1+2n1, where n=0,±1,±2,±3,… . (4.57) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 203 For the case n=0, we have: 1=20=21+21. (4.58) In the book [46] the following “dynamic” model of the Dichotomy Princi ple is given in the form of the infinite division of the “Unit” (“The Whole”) according to the “dichotomy” relations (4.57) and (4.58): 1=20 =21 +21 21 =22 +22 22 =23 +23 ∞ (4.59) 1= 20 = 21 + 22 + 23 + 24 = ... = ∑ 2i . i =1 4.5.2. Classical Golden Section Principle The Golden Section Principle that came to us from Pythagoras, Plato, and Euclid is based on the following fundamental property that connects the ad jacent powers of the golden mean τ = 1 + 5 2 : ( ) τ =τ +τ , where n=0,±1,±2,±3,… . For the case n=1, the identity (4.60) takes the following form: n n1 n2 (4.60) (4.61) 1=τ0=τ1+τ2. Using the golden identities (4.60) and (4.81), Shevelev developed [46] the following “dynamic” model of the Golden Section Principle: 1 = τ0 = τ −1 + τ −2 τ −2 = τ −3 + τ −4 τ −4 = τ −5 + τ −6 1 = τ0 = τ −1 + τ −3 + τ −5 + τ −7 + ... = ∞ ∑τ ( − 2 i −1) . (4.62) i =1 Note that the Dichotomy Principle (4.59) and the Golden Section Principle (4.62) have a great number of applications in nature, science and mathemat ics (binary number system, numerical methods of the algebraic equation solu tions, self division and so on). 4.5.3. The Generalized Principle of the Golden Section By dividing all terms of the identity (4.48) by τnp , we obtain the following identity: 1 = τ0p = τ −p1 + τ −p p −1 . (4.63) Alexey Stakhov THE MATHEMATICS OF HARMONY 204 Using (4.48) and (4.63), we can construct the following “dynamic” model of the “Unit” decomposition according to the Golden pProportion: 1 = τ0p = τ −p1 + τ p( ) − p +1 − p +1 −1 −2 p +1 τ p( ) = τ p( ) + τ p ( ) −2 p +1 −1 −2 p +1 −3 p +1 τp ( ) = τp ( ) + τp ( ) − p +1 1 = τ0p = τ −p1 + τ p( − p +1)−1 + τp ( −2 p +1) −1 + τp ( −3 p +1)−1 + ... = ∞ ∑ − i −1 p +1 −1 τ p( )( ) . (4.64) i =1 The main result of the above consideration is to find more general principle of the “Unit” division that is given by the following identity: 1 = τ −p1 + τ −p( p +1) = ∞ ∑τ − ( i −1)( p +1) −1 , p i =1 (4.65) where τp is the golden pproportion, p=0,1,2,3,… . It is clear that this general principle Generalized Principle of the Golden Section includes the Dichotomy Principle (4.59) and the classical Golden Section Principle (4.62) as special cases for p=0 and p=1, respectively. 4.6. A Generalization of Euclid’s Theorem II. 11 4.6.1. A Generalization of Euclid’s Theorem II.11 for the Case p=2 As we mentioned above, the golden psections that are given by the pro portion (4.44) is a generalization of the classical golden section that is given by (1.3). However, in Euclid’s Theorem II.11 the DEMR is formulated in the form (1.2). We can try to represent the proportion (4.44) in the form (1.2). We start from the partial case p=2. For this case the proportion (4.44) takes the following form: 2 CB AB (4.66) . = AC CB Let us denote the lengths of the line segments АВ, АС and СВ in (4.66) as follows: AB=a, CB=b, AC=c. Then, it can be represented in the form: (4.67) a2×c=b3. We can give the following geometric interpretation of the equality (4.67). The righthand part of the equality (4.67) can be interpreted as the volume of a cube with the side equal to b, that is, to the length of the larger segment CB Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 205 that arises at the division of a line segment АВ in the golden 2proportion (4.66). The lefthand part of the equality (4.67) can be interpreted as the volume of a rectangular parallelepiped. This parallelepiped has a square at its base with sides equal to а, that is, to the length of the initial segment АВ. The height of the rectangular parallelepiped is equal to с, that is, to the length of the smaller seg ment АС in the proportion (4.66). Then, taking into consideration (4.66) and (4.67), we can formulate a new geometric problem of the division of a line in the golden 2section that is a generalization of Euclid’s DEMR. Generalization of DEMR (a Division in the Golden 2Proportion). Di vide the given line АВ at point C into two segments, the smaller segment АС and the larger segment СВ, so that a volume of the cube with the side equal to the larger segment CB is equal to the volume of a rectangular parallelepiped with a base, which is a square with sides equal to the initial line АВ, and with the height equal to the smaller segment AC. 4.6.2. Euclid’s Rectangular Parallelepiped The rectangular parallelepiped appearing in the above problem consists of 6 faces (Fig. 4.7). The top and bottom faces are squares with sides equal to the length of the initial segment a; the lateral faces are the rectangles with sides equal to a and c. These rectangles are similar to Euclid’s rectangle in Fig. 4.7 where the ratio of its sides a:c for the given case is equal to the square of the golden 2proportion τ2, that is, a = τ22 . (4.68) c a We will name this geometric figure Euclid’s Rect angular Parallelepiped. Thus, according to (4.68) the ratio of the side of its base to its height in Euclid’s Rect c angular Parallelepiped is equal to the square of the golden 2proportion τ2; here, according to (4.67) its volume is a cube of the length of the larger segment in the proportion (4.66). Figure 4.7. If the initial segment AB is a unit segment (AB=1), Euclid’s rectangular parallelepiped then the equality (4.67) takes the following form: c=b3. (4.69) Then, we can formulate the following geometric problem of the division of the unit segment in the golden 2proportion. Alexey Stakhov THE MATHEMATICS OF HARMONY 206 A Problem of the Division of the Unit Segment in the Golden 2Propor tion. Divide a unit segment into two unequal segments such that the smaller segment’s length is equal to the cube of the larger segment’s length. Note that the formulated problem expresses the following property of the golden 2proportion: 1 = τ2−1 + τ2−3 = 0.6823 + 0.3177. (4.70) 4.6.3. The General Case of p For the general case of р, we represent (4.44) in the following form: a ×c=bp+1. (4.71) By using geometric language, we can interpret the equality (4.71) as fol lows. The righthand part of the equality (4.71) is a volume of a hypercube in the (p+1)dimensional space with the side equal to the length b of the larger segment of the division of a line in the golden рproportion. The lefthand part of the equality (4.71) is a volume of Euclid’s HyperRectangular Parallel epiped in the (p+1)dimensional space; here, the р sides are equal to the length а of the initial segment at the division of a line in the golden рproportion, and the (p+1)th side (its “height”) is equal to the length с of the smaller segment at the division of a line in the golden рproportion. It is clear that (4.71) expresses a generalized Euclidean problem of the division in extreme and mean ratio. We can consider this problem for the case of the unit segment (a=1). Then, the equality (4.71) takes the following form: p (4.72) c=bp+1. Taking into consideration (4.72), we can formulate the following problem. A Problem of the Division of the Unit Segment in the Golden pPro portion. For a given p=0,1,2,3,… divide a unit segment into two unequal seg ments in such proportion that the smaller segment’s length is equal to the (p+1)th degree of the larger segment’s length. 4.7. The Roots of the Generalized Golden Algebraic Equations 4.7.1. Algebraic Equations As is well known, algebraic equations have the following general properties, which can be used by us before finding their roots: Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 207 1. Algebraic equations of the nth degree have n roots. In special cases it can appear, that some roots are repeated some times (multiple roots); hence, the number of various roots can be less than n. 2. Descartes’ “rule of signs”: the algebraic equation has no more positive roots than the number of sign changes in a series of its factors. Often, it is not so important to calculate the roots, it is more important to understand the character of these roots. “The rule of signs” helps to solve this problem. For example, the equation x54x2=0 has only one positive root because the series of their factors 1,4,2 have only one change of a sign. 3. In the equations with real factors, the complex roots can appear only by pairs: alongside with the root a+bi the complex number abi is always a root of the same equation. 4. If xi (i=1,2,3,…,n) is one of the roots of the algebraic equation (4.73) a0xn+ a1xn1+…+an=0, then it is easy to prove, that a polynomial that stands in the lefthand part of Eq. (4.73), is divided by the binom (xxi) without remainder. It is easy to prove that any polynomial of nth degree can be represented as a prod uct of the n multipliers of the 1st degree of the kind (xxi), that is, (4.74) a0xn+ a1xn1+…+an=(xx1)(xx2) … (xxn). This theorem (4.74) is sometimes named the Basic Theorem of Algebra. 4.7.2. Properties of the Roots of the Generalized Golden Algebraic Equations Once again, consider the characteristic equation (4.42). By using the above rules 14, we can prove the following properties of the roots of Eq. (4.42): 1. According to Descartes’ “Rule of Signs,” Eq. (4.42) has the only positive root τp. This root expresses some important property of Pascal Triangle. 2. As Eq. (4.42) has the degree (p+1), this means that Eq. (4.42) has (р+1) roots x1, x2, …, xp, xp+1. Further, without loss of generality, we suppose that the root x1 always coincides with the golden рproportion τp, that is, x1=τp. 3. As Eq. (4.42) has only the real factors 1, 1 and 1, this means that all complex roots of Eq. (4.42) appear in pairs, that is, each complex root a+bi appears always together with the root abi, which is the complex conjugate to the root a+bi. 4. By using (4.74), we can represent Eq. (4.42) by its roots as follows: xp+1xp1 =(xx1)(xx2) … (xxp)(xxp+1). (4.75) The general identity below for the roots x1, x2, …, xp, xp+1 comes from (4.42): THE MATHEMATICS OF HARMONY Alexey Stakhov 208 (4.76) x kn = x kn −1 + x kn − p −1 = x k × x kn −1 , where n=0,±1,±2,±3,… and xk (k=1,2,3,…, p+1) is a root of Eq. (4.42). Note that the identity (4.48) is a partial case of the identity (4.76) for the case k=1. Theorem 4.1. For a given integer p>0, the following correlations for the roots of the characteristic equation xp+1xp1=0 are valid: x1 + x 2 + x 3 + ... + x p + x p +1 = 1 (4.77) ( x x + x x + ... + x x ) + ( x x + ... + x x ) + ... + x x = 0 ( x x x + x x x + ... + x x x ) + ( x x x + ... + x x x ) + ... + x 1 2 1 3 1 p +1 2 3 2 p +1 p p+1 p−1 x p x p +1 = 0 ... x1 x2 " x p−2 x p−1 x p + x1 x 3 x 4 " x p−1 x p x p+1 + ... + x2 x 3 x 4 " x p−1 x p x p+1 = 0 (4.78) x1 x2 x 3 " x p−1 x p x p+1 = (−1) . (4.79) 1 2 3 1 3 4 1 p +1 p 2 3 4 2 p p +1 p Proof. Consider the representation of Eq. (4.42) in the form (4.75). If we remove the parentheses in (4.75), then, for the even p=2k we can write: x p+1 − x p −1 = ( x − x1 )( x − x2 )"( x − x p+1 ) = x p+1 −( x1 + x2 + ... + x p+1 ) x p +( x1 x2 + ... + x1 x p+1 + x2 x 3 + ... + x2 x p+1 + ... + x p−1 x p+1 + x p x p+1 ) x p−1 −( x1 x2 x 3 + ... + x1 x p x p+1 + x2 x 3 x 4 + ... + x2 x p x p+1 + ... + x p−1 x p x p+1 ) x p−2 +( x1 x2 x 3 x 4 + x1 x2 x 3 x5 + ... + x p−2 x p−1 x p x p+1 ) x p−3 ... +( x1 x2 " x p−1 x p + x1 x3 " x p x p+1 + ... + x2 x3 " x p x p+1 ) x − x1 x2 " x p x p+1 = 0. (4.80) The following outcomes follow from the comparison of (4.42) and (4.80): x1 + x2 + ... + x p + x p+1 = 1 +( x x + x x )+ x x ( x x + ... + x x ) + ( x x + ... + x x )+ ...+ ( x x x + ... + x x x ) + ( x x x + .... + x x x ) + ... + x x x = 0 1 2 1 2 1 3 p +1 1 p 2 3 p +1 2 2 3 p−1 p +1 2 4 p p p +1 p−1 p +1 p−1 p p p +1 =0 p +1 (4.81) ... x1 x2 " x p−2 x p−1 x p + x1 x3 x4 " x p−1 x p x p+1 + ... + x2 x 3 x 4 " x p−1 x p x p+1 = 0 x1 x2 " x p x p+1 = 1. For the odd p=2k+1, we have: x p+1 − x p −1 = ( x − x1 )( x − x2 )"( x − x p+1 ) = x p+1 −( x1 + x2 + …+ x p+1 ) x p +( x1 x2 + … + x1 x p+1 + x2 x3 + … + x2 x p+1 + …+ x p−1 x p + x p−1 x p+1 + x p x p+1 ) x p−1 −( x1 x2 x 3 + … + x1 x p x p+1 + x2 x 3 x 4 + …+ x2 x p x p+1 + …+ x p−1 x p x p+1 ) x p−2 +( x1 x2 x 3 x 4 + x1 x2 x 3 x5 + …+ x p−2 x p−1 x p x p+1 ) x p−3 ... −( x1 x2 x 3 " x p−2 x p−1 x p + … + x2 x3 x4 " x p−1 x p x p+1 ) x + x1 x2 x3 " x p−1 x p x p+1 = 0. (4.82) The following outcomes follow from the comparison of (4.42) and (4.82): x1 + x2 + ... + x p + x p+1 = 1 +( x x + x x )+ x x ( x x + ... + x x ) + ( x x + ... + x x )+ ...+ ( x x x + ... + x x x ) + ( x x x + .... + x x x ) + ... + x x x = 0 1 2 1 2 1 3 p +1 1 p 2 p +1 3 2 2 3 4 p−1 p +1 2 p p +1 p p−1 p +1 p−1 p p p +1 =0 p +1 ... x1 x2 " x p−2 x p−1 x p + x1 x3 x4 " x p−1 x p x p+1 + ... + x2 x 3 x 4 " x p−1 x p x p+1 = 0 x1 x2 " x p x p+1 = −1. (4.83) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 209 The outcomes (4.81) and (4.83) prove Theorem 4.1. It is evident from (4.77) that the sum of the roots of Eq. (4.42) is identi cally equal to 1. The expression (4.78) gives the values for every possible sum of the roots of Eq. (4.42) taken by two, three, ..., or р roots from the (p+1) roots of Eq. (4.42). According to (4.78), each of these sums is identically equal to zero! At last, the expression (4.79) gives the value of the product of all roots of Eq. (4.42). According to (4.79) this product is equal to 1 (for the even р) or 1 (for the odd р). Theorem 4.2. For a given integer p=1,2,3,… and for the condition when k takes its values from the set {1,2,3,…,p}, the following identity is valid for the roots of the characteristic equation xp+1xp1=0: ( x + x + x + ... + x ) = x + x + x + ... + x k 1 2 3 p +1 k 1 k 2 k 3 k p +1 = 1. (4.84) Proof. Consider the expression: (x1+ x2+…+ xp+ xp+1)k, where k takes its values from (4.86) k∈{1,2,…,p}. Taking into consideration the identity (4.77), we can write: (4.85) (4.86) (4.87) (x1+ x2+…+ xp+ xp+1)k =1k=1. Consider a partial case of the expression (4.87) for the case p=1. Taking into consideration the condition (4.86) for the case p=1 the expression (4.85) can take the only form: (4.88) x1+x2. According to the property (4.87), we can write: (4.89) x1+x2 = 1 that satisfies the expression (4.84). Now, consider the case p=2. Taking into consideration the condition (4.86), for the case p=2 the expression (4.85) can take only two different forms: x1+x2+x3 and (4.90) (4.91) (x1+x2+x3)2. Taking into consideration the identity (4.87), we can write for the case (4.90) the following identity: x1+x2+x2 = 1. Now, consider the case (4.91). Representing (4.91) in the form (4.92) [(x1+x2)+ x3]2, (4.93) Alexey Stakhov THE MATHEMATICS OF HARMONY 210 we can write: ( x1 + x 2 + x 3 )2 = [( x1 + x 2 ) + x 3 ] = x12 + x 22 + x 32 + 2 ( x1 x 2 + x1 x 3 + x 2 x 3 ) . 2 (4.94) If we take into consideration the properties (4.78) and (4.84), we can re write the expression (4.94) as follows: ( x1 + x2 + x3 )2 = x12 + x22 + x32 = 1. (4.95) By analogy, we can write 3 possible expressions like (4.85) for the case p=3: x1+x2+x3+x4, (4.96) (x1+x2+x3+x4)2, (4.97) (x1+x2+x3+x4)3. (4.98) If we fulfill simple transformations of the expressions (4.96)(4.98) and take into consideration the properties (4.77) and (4.84), we can write 3 iden tities for the case p=3: x1+x2+x3+x4 = 1 (4.99) ( x1 + x2 + x3 + x4 )2 = x12 + x22 + x32 + x42 = 1 3 ( x1 + x2 + x3 + x4 ) = x13 + x23 + x33 + x43 = 1. (4.100) (4.101) In the generalized case, for the proof of Theorem 4.2 for arbitrary p we can use the socalled Multinomial Theorem [35], which is a generalization of the binomial theorem (4.4). For any positive integer m and any nonnegative in teger n, the multinomial formula is the following: n k1 k2 k3 km ( x1 + x2 + x3 + ... + x m )n = ∑ x1 x2 x 3 ... x m . (4.102) k1 , k2 , k3 ,..., km k1 , k2 , k3 ,..., km The summation is taken over all sequences of the nonnegative integer in m dexes k1 through km such that ∑ ki = n . The numbers i =1 n n! k , k , k ,..., k = (4.103) k1 ! k2 ! k3 !...km ! m 1 2 3 are called Multinomial Coefficients. Note that the Binomial Theorem (4.4) and the Binomial Coefficients (4.2) are special cases (m=2) of the Multinomial Formula (4.102) and the Multino mial Coefficients (4.103), respectively. In the general case of p, the expression (4.85) can be factorized if we use the Multinomial Formula (4.102). As is known [35], for a given k, the Multino mial Formula (4.102) will include in itself the sum of all kth powers of Eq. (4.42) that are taken with the coefficient of 1, that is, x1k + x2k + x3k + ... + x pk + x pk +1 , (4.104) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 211 and the sum of the products of every possible combination of two (k=2), three (k=3) or the k roots of Eq. (4.42) that are taken with the factors known as Multinomial Coefficients (4.103) [35]. According to Theorem 4.1 all these sums are identically equal to zero. And then taking into consideration (4.86), we can write the general identity (4.84). 4.7.3. Some Corollaries of Theorems 4.1 and 4.2 Now, consider some corollaries of Theorems 4.1 and 4.2 for the different val ues of р. It is well known that for the case p=1 the golden algebraic equation (4.43) has two real roots: 1+ 5 1 1− 5 . and x2 = − = τ 2 2 Hence, from the above we can obtain the following identities correspond ing to Theorem 4.1 for the roots x1 and x2: x1 = τ = (4.105) x1+x2=1; x1×x2=1. For the case p=2, the golden algebraic equation (4.42) takes the following form: x3=x2+1. (4.106) Equation (4.106) has three roots one real root x1 and two complex con jugate roots x2 and x3 that are given below: h 2 1 x1 = + + = 1.4655712319... (4.107) 6 3h 3 x2 = − 1 1 3 h 2 h − + −i ( − ) = −0.233... − (0.793...)i 12 3h 3 2 6 3h (4.108) x3 = − h 1 1 3 h 2 − + +i ( − ) = −0.233... + (0.793...)i , 12 3h 3 2 6 3h (4.109) where (4.110) h = 3 116 + 12 93 . Using direct substitution for the roots x1, x2 and x3, it is easy to prove the following identities corresponding to Theorems 4.1 and 4.2 for the case p=2: x1 + x 2 + x 3 = 1, x1 x 2 + x1 x 3 + x 2 x 3 = 0, x1 × x 2 × x 3 = 1, x12 + x 22 + x 32 = 1. Alexey Stakhov THE MATHEMATICS OF HARMONY 212 4.8. The Generalized Golden Algebraic Equations of Higher Degrees 4.8.1. The Case of p=1 In Chapter 1 we developed a theory of the golden algebraic equations of higher degrees. We proved that for the simplest golden algebraic equation (4.43) – which is a partial case (p=1) of the characteristic equation (4.42) there is an infinite number of the characteristic equations with the degree greater than 2 that have ( ) the golden mean τ = 1 + 5 2 as their general root. In their general form these golden equations are given by the expression: xn=Fnx2Fn2= FnxFn1, (4.111) where Fn, Fn1, Fn2 are Fibonacci numbers. As we mentioned in Chapter 1, the equation x4=3x+2 which is a partial case of the general equation (4.111) corresponding to n=4 describes the energy state of the butadiene molecule, a valuable chemical substance, which is used in the production of rubber. This fact at once places our interest in the golden equations (4.111), because they, probably, in general, can describe the energy conditions of the molecules of other chemical substances. In this connection it is of great interest to study the characteristic equa tions given by (4.42) and the algebraic equations following from them with degrees more than p+1 which have the golden pproportions τp as their gen eral roots. Let us demonstrate our approach for the partial cases p=2,3. 4.8.2. The Case of p=2 For the case of p=2 Eq. (4.42) is reduced to the algebraic equation of the third degree: x3=x2+1. Multiplying repeatedly all terms of the equation x3=x2+1 by x, we can find the following equality for the case of p=2: (4.112) xn=xn1+ xn3, where n=3, 4, 5,… . Using the equality (4.112) and (4.42), we derive the following algebraic equations that have the golden 2proportion τ2 as a root: x4 = x3 + x = x2 + x + 1 x 5 = x 4 + x 2 = 2x 2 + x + 1 x 6 = x 5 + x 3 = 3x 2 + x + 2 ... x n = x n −1 + x n −3 = F2 ( n − 1) x 2 + F2 ( n − 3) x + F2 ( n − 2) , (4.113) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 213 where F2(n1), F2(n2), F2(n3) are the Fibonacci 2numbers that are given by the recursive relation (4.14) at the seeds (4.15). As the golden 2proportion τ2 is the root of each of the equations that are given by (4.113), the following general identity, which connects the golden 2 proportion τ2 with the Fibonacci 2numbers F2(n), follows from (4.113): τ2n = τ2n −1 + τnn −− 33 = F2 ( n − 1) τ22 + F2 ( n − 3 ) τ2 + F2 ( n − 2) . (4.114) 4.8.3. The Case of p=3 Consider the characteristic equation (4.42) for the case of p=3: x =x3+1. (4.115) If we use the reasoning similar to the case of p=2, we can obtain the following algebraic equations that have the golden 3proportion τ3 as their general root: 4 x5 = x4 + x = x3 + x + 1 x6 = x5 + x2 = x3 + x2 + x + 1 x 7 = x 6 + x 3 = 2x 3 + x 2 + x + 1 (4.116) ... n n −1 n −4 3 2 x = x + x = F3 ( n − 2 ) x + F3 ( n − 5 ) x + F3 ( n − 4 ) x + F3 ( n − 3 ) , where F3(n2), F3(n3), F3(n4), F3(n5) are the Fibonacci 3numbers given by the recursive relation F3(n)=F3(n1)+F3(n4) at the seeds: F3(1)=F3(2)=F3(3)=F3(4)=1. As the golden 3proportion τ3 is the root of any of the equations (4.116), the following identity, which connects the golden 3proportion τ3 with the Fibonacci 3numbers F3(n), follows from (4.116): τ3n = τ3n −1 + τ3n − 4 = F3 ( n − 2) τ33 + F3 ( n − 5) τ23 + F3 ( n − 4 ) τ3 + F3 ( n − 3 ) . (4.117) 4.8.4. The General Case For the general case of р, we can write the following equality, which can be obtained from the algebraic equation (4.42): xn=xn1+ xnp1, (4.118) where n=p+1, p+2, p+3, … . Using Eqs. (4.42) and (4.118), we can obtain the following formula for the generalized characteristic equations that have the golden pproportion τp as their root: x n = Fp ( n − p + 1) x p + p −1 ∑ F (n − p − t ) x , t p t =0 (4.119) Alexey Stakhov THE MATHEMATICS OF HARMONY 214 where n=p+1, p+2, p+3, …; Fp(np+1) and Fp(npt) are the Fibonacci рnum bers given by the recursive relation (4.18) at the seeds (4.19). It is clear that, for a given p>0, the formula (4.119) sets an infinite num ber of generalized characteristic equations with the general root τp. The fol lowing general identity that comes from (4.119) connects the degrees of the golden рproportion τp with the Fibonacci рnumbers Fp(n): τnp = Fp ( n − p + 1) τ pp + p −1 ∑ F (n − p − t ) τ , p t p t =0 (4.120) where n=p+1, p+2, p+3, … . Concluding this Section we may note that the above algebraic equations that are given by (4.111), (4.113), (4.116), and (4.119) are unusual algebraic equations. First of all, they follow from Pascal Triangle and express some im portant mathematical properties. Besides, they describe some harmonious chemical and physical structures and we may expect their application for sim ulation of many physical and chemical processes and structures. 4.9. The Generalized Binet Formula for the Fibonacci рNumbers 4.9.1. A General Approach to the Synthesis of the Generalized Binet Formulas In Chapter 2 we established the Binet formulas (2.62) and (2.63). These formulas are the representations of the “extended” Fibonacci and Lucas num bers by the golden mean τ. In this Section we try to develop a general approach to the synthesis of Binet formulas based on the representation of the “extended” Fibonacci pnumbers by the roots of the characteristic equation (4.42). Our main hypothesis is the following. For a given p>0, we can represent the Binet formula for the Fibonacci рnumbers as follows: n n n Fp ( n ) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) , (4.121) where x1, x2, …, xp+1 are the roots of Eq. (4.42), and k1, k2, …, kp+1 are constant coefficients that depend on the initial elements of the Fibonacci pseries. It follows from (4.29) that Fp(0)=0 for any p>0. Therefore, we can calculate the Fibonacci рnumbers according to the recursive relation (4.18) at the seeds: Fp(0)=0, Fp(1)=Fp(2)=Fp(p)=1. (4.122) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 215 Taking into consideration (4.121) and (4.122), we can write the following system of algebraic equations: Fp ( 0 ) = k1 + k2 + ... + kp +1 = 0 Fp (1) = k1 x1 + k2 x2 + ... + kp+1 = 1 Fp ( 2) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) = 1 ... p p p Fp ( p ) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) = 1. 2 2 2 (4.123) Solving the system of the equations (4.123), we obtain the numerical val ues of the coefficients k1, k2, …, kp+1 for the different values of p. 4.9.2. A Derivation of Binet Formulas for the Classical Fibonacci and Lucas Numbers We can use the general formula (4.121) to obtain Binet formulas for the case p=1. For this case, the characteristic equation (4.42) is reduced to Eq. (4.43), which has two roots x1=τ and x2=1/τ, where τ = 1 + 5 2. Therefore, formula (4.121), for the case of p=1, takes the following form: n n 1 F1 ( n ) = k1 ( τ ) + k2 − . (4.124) τ It is clear that for the case of p=1 the system of algebraic equations (4.123) is F1 ( 0 ) = k1 + k2 = 0 (4.125) F1 (1) = k1τ + k2 ( − 1 τ ) = 1. By solving the system (4.125), we obtain: k1 = 1 5 and k1 = − 1 5 . If we substitute k1 and k2 into (4.124), we obtain the wellknown Binet formula for the classical Fibonacci numbers in the form: n τn − ( − 1 τ ) . F1 (n) = (4.126) 5 If we assume k1=k2=1 in (4.124), we obtain the Binet formula for the clas sical Lucas numbers: ( ) (4.127) L1 (n) = τn + ( − 1 τ ) . This formula generates the Lucas series at the seeds L1(0)=2 and L1(1)=1: n 2, 1, 3, 4, 7, 11, 18, 29, … . (4.128) Let us note one important fact regarding the seed F1(0)=0. It follows di rectly from the Binet formula (4.124). In fact, according to (4.124) we have the following result for the case of n=0: Alexey Stakhov THE MATHEMATICS OF HARMONY 216 (4.129) F1 ( 0 ) = k1τ0 + k2 ( − 1 τ ) = k1 + k2 = 1 5 − 1 5 = 0. These simple calculations show that the seed F1 ( 0 ) = 0 comes from the fact that the sum of the coefficients k1 and k2 in the expression (4.125) is also equal to zero. 0 4.9.3. Binet Formula for the Fibonacci 2Numbers Let us give p=2 and use the above approach for obtaining Binet formula for the Fibonacci 2numbers. For the case p=2, the recursive relation for the Fibonacci 2numbers and the characteristic equation (4.42) take the follow ing forms, respectively: F2(n)=F2(n1)+F2(n3) (4.130) F2(0)=0, F2(1)=F2(2)=1 (4.131) x3=x2+1. (4.132) Equation (4.132) has three roots real (positive) root x1=τ2 given by (4.107) and two complexconjugate roots x2 and x3 (4.108) and (4.109). We recall that the real root x1 of the algebraic equation (4.132) is an irra tional number that is equal to the golden 2proportion (x1=τ2). The number h given by (4.110) is also irrational; hence, the roots x2 and x3 are complex num bers with irrational real parts. For the case p=2, the formula (4.121) and the system (4.123) take the following forms, respectively: n n n F2 ( n ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) (4.133) F2 ( 0 ) = k1 + k2 + k3 = 0 F2 (1) = k1 x1 + k2 x2 + k3 x 3 = 1 F2 ( 2) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) = 1. 2 2 2 (4.134) Solving the system (4.134), we obtain: k1 = 2h ( h + 2 ) ( h + 8) (4.135) 3 − ( h + 2) + i 3 ( h − 2) h k2 = 3 h +8 ( ) (4.136) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 217 − ( h + 2) − i 3 ( h − 2) h , k3 = h3 + 8 ( ) (4.137) where h is given by (4.110). Note that the coefficient (4.135) is an irrational number because the num ber h given by (4.110) is irrational. Also the coefficients (4.136) and (4.137) are complexconjugate numbers with irrational real parts. Substituting (4.107)(4.110) and (4.135)(4.137) into the expression (4.133), we can write the following Binet formula for the Fibonacci 2 numbers, which is a generalization of the Binet formula (4.126) for the classi cal Fibonacci numbers : F2 ( n ) = 2h ( h + 2 ) h 2 1 + + h3 + 8 6 3h 3 n n − ( h + 2) + i 3 ( h − 2) h h 1 1 3 h 2 + + −i − − − 2 6 3h h3 + 8 12 3h 3 n − ( h + 2 ) − i 3 ( h − 2 ) h h 1 1 3 h 2 + +i + − − 6 − 3h . 12 3 3 2 h h3 + 8 (4.138) It seems incredible at first sight, that the formula (4.138) that is very complicated combination of complex numbers with irrational coeffi cients represents the integer Fibonacci 2numbers F 2(n) for any integer n=0, ±1, ±2, ±3, ... . 4.9.4. Binet Formula for the Fibonacci 3Numbers For the case of p=3, the recursive relation for the Fibonacci 3numbers and the characteristic equation (4.42) take the following forms, respectively: F3(n)=F3(n1)+F3(n4) (4.139) F3(0)=0, F3(1)=F3(2)=F3(3)=1 (4.140) x4=x3+1. (4.141) Equation (4.141) has four roots two real roots, x1 and x2, and two com plexconjugate roots, x3 and x4. The roots of Eq. (4.141) are irrational and complex numbers, which have a very complex symbolic representation; there fore, we can use their approximate numerical values: x1=1.380; x2=0.819; x3=0.219+0.914i; x4=0.2190.914i. (4.142) Alexey Stakhov THE MATHEMATICS OF HARMONY 218 It is important to note that the root x1 of Eq. (4.141) is the golden 3propor tion (x1=τ3=1.380). The formulas (4.122) for the Fibonacci 3numbers and the system of the algebraic equations (4.123) take the following forms, respectively: F3 ( n ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) + k4 ( x4 ) n n n n (4.143) F3 ( 0 ) = k1 + k2 + k3 + k4 = 0 F3 (1) = k1 x1 + k2 x2 + k3 x3 + k4 x4 = 1 F3 ( 2) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) + k4 ( x4 ) = 1 2 2 2 2 (4.144) F3 ( 3 ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) + k4 ( x4 ) = 1. 3 3 3 3 Solving the system (4.144), we obtain the following numerical coefficients: (4.145) k1=0.3969; k2=0.1592; k3=0.11880.2045i; k4=0.1188+0.2045i. Hence, using (4.142) and (4.145), we can represent Binet formula (4.143) for the Fibonacci 3numbers in the following numerical form: F3 ( n ) = 0.3969 (1.38 ) − 0.1592 ( −0.819 ) n + ( −0.1188 − 0.2045i ) ( 0.219 + 0.914i ) n + ( −0.1188 + 0.2045i ) ( 0.219 − 0.914i ) . n n (4.146) 4.9.5. Binet Formulas for the Fibonacci 4Numbers For the case of p=4, the recursive relation for the Fibonacci 4numbers and the characteristic equation (4.42) take the following forms, respectively: F4(n)= F4(n1)+F4(n5) (4.147) F4(0)=0, F4(1)=F4(2)= F4(3)= F4(4)=1 (4.148) x5=x4+1. (4.149) Equation (4.141) has five roots one real root x1 that coincides with the gold en 4proportion τ4 and two pairs of complexconjugate roots x2, x3 and x4, x5. They all can be represented in the following analytical and numerical form: x1 = h 2 1 i 3 1 i 3 + = 1.3247, x2 = − = 0.5 − 0.866i, x 3 = + = 0.5 + 0.866i 6 h 2 2 2 2 h 1 i 3 h 2 x4 = − + − − = −0.6623 − 0.5623i 12 h 2 6 h h 1 i 3 h 2 x5 = − + + − = −0.6623 + 0.5623i, 12 h 2 6 h (4.150) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 219 where h = 3 108 + 12 69 . For this case, the formula (4.121) for the Fibonacci 4numbers and the system of the algebraic equations (4.123) take the following forms: n n n n n F4 ( n ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) + k4 ( x4 ) + k5 ( x5 ) (4.151) F4 ( 0 ) = k1 + k2 + k3 + k4 + k5 = 0 F4 (1) = k1 x1 + k2 x2 + k3 x 3 + k4 x 4 + k5 x5 = 1 F4 ( 2 ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) + k4 ( x4 ) + k5 ( x5 ) = 1 2 2 2 2 2 F4 ( 3 ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x 3 ) + k4 ( x 4 ) + k5 ( x5 ) = 1 3 3 3 3 3 (4.152) F4 ( 4 ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x 3 ) + k4 ( x 4 ) + k5 ( x5 ) = 1. Solving the system (4.152), we obtain the numerical values of the coefficients: 4 4 4 4 k1 = 0.380; k2 = −0.171 + 0.206i; k3 = −0.071 − 0.206i; k4 = −0.119 + 0.046i; k5 = −0.119 − 0.046i. 4 (4.153) Using (4.150) and (4.153), we can represent the Binet formula for the Fibonacci 4numbers (4.151) in numerical form. 4.9.6. A General Case of p In the general case, Binet formula for the Fibonacci рnumbers has the form (4.121). The coefficients k1, k2, …, kp+1 in the formula (4.121) are the solu tions of the system (4.123). This outcome can be formulated as the following theorem. Theorem 4.3 (Generalized Binet Formula for the Fibonacci p Numbers). For a given integer p>0, any Fibonacci рnumber Fp(n) (n=0, ±1, ±2, ±3, …) given by the recursive relation Fp(n)=Fp(n1)+Fp(np1) at the seeds Fp(0)=0, Fp(1)=Fp(2)= … =Fp(p)=1 can be represented in the following ana lytical form: Fp ( n ) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) , where x1, x2, …, xp+1 are the roots of the characteristic equation xp+1=xp+1 and k1, k2, …, kp+1 are constant coefficients that are the solutions of the system of the algebraic equations: n n n Fp ( 0 ) = k1 + k2 + ... + kp +1 = 0 Fp (1) = k1 x1 + k2 x2 + ... + kp +1 x p +1 = 1 ( ) 2 Fp ( 2 ) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 x p +1 = 1 ... p p p Fp ( p ) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 x p +1 = 1. 2 2 ( ) Alexey Stakhov THE MATHEMATICS OF HARMONY 220 Proof. Represent the Fibonacci рnumber Fp(p+1) in the analytical form according to (4.121): p +1 p +1 p +1 Fp ( p + 1) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) . (4.154) Using the identity (4.76), we can represent the expression (4.154) in the following form: p p p F p ( p + 1) = k1 ( x1 ) + k2 ( x 2 ) + ... + kp +1 x p +1 0 0 0 (4.155) + k1 ( x1 ) + k2 ( x 2 ) + ... + k p +1 x p +1 . Then, using the definition (4.121), we can write: ( ( ) ) Fp(p+1)= Fp(p)+Fp(0). (4.156) This means that the recursive relation (4.18) is valid for the Fibonacci pnumber Fp(p+1), that is, the analytical formula (4.154) represents the Fibonacci pnumber Fp(p+1). Furthermore, applying the formula (4.121) for the Fibonacci рnumbers Fp(p+2), Fp(p+3), …, Fp(n), … and using the identity (4.76), it is easy to prove that the analytical formula (4.121) represents all Fibonacci pnumbers for the positive values of n. Let us prove that Eq. (4.121) is valid for the negative values of n=1, 2, 3, … . In order to do this, we consider the formula (4.121) for the case n=1: −1 −1 −1 Fp ( −1) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) . (4.157) We can rewrite the identity (4.76) in the following form: x kn − p −1 = x kp − x kn −1 . (4.158) For the case n=p, the identity (4.158) takes the following form: x k−1 = x kp − x kp −1 . (4.159) Using the identity (4.159), we can rewrite (4.157) as follows: ( ) ( ) p p p F p ( −1) = k1 ( x1 ) + k2 ( x 2 ) + ... + kp +1 x p +1 p −1 p −1 p −1 (4.160) . − k1 ( x1 ) + k2 ( x 2 ) + ... + kp +1 x p +1 Hence, using the general formula (4.121), we can see that the formula (4.160) is equivalent to Fp(1)= Fp(p)+Fp(p1)=11=0, that is, the formula (4.160) represents the Fibonacci рnumber Fp(1)=0. Furthermore, considering the formula (4.121) for the negative values of n=2, 3, 4, … and using (4.158), it is easy to prove that the formula (4.121) is valid for all negative values of n. Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 221 4.10. The Generalized Lucas pNumbers 4.10.1. Binet Formula for the Lucas pNumbers Next let us generalize the Binet formula for the classical Lucas numbers that are given by (4.127). We recall that the formula (4.127) can be obtained from the Binet formula for Fibonacci numbers that are given by (4.126), provided we assume in it that k1=k2. We can use this approach for the introduction of a new class of the recursive sequences. If we assume that k1=k2=…=kp+1=1 in the for mula (4.121), we can obtain the following formula: Lp ( n ) = ( x1 ) + ( x2 ) + ... + ( x p +1 ) . (4.161) Note that for the case p=1 this formula is reduced to the Binet formula for the classical Lucas numbers that are given by (4.127). Let us prove that this formula represents a new class of recursive numerical sequences called Lucas pnumbers. They are given by the following recursive relation: n n n Lp(n)= Lp(n1)+Lp(np1) at the seeds: (4.162) Lp(0)=p+1 (4.163) Lp(1)=Lp(2)=…=Lp(p)=1. (4.164) In fact, for the case n=0, we can write the formula (4.161) as follows: Lp ( 0 ) = ( x1 ) + ( x2 ) + ... + ( x p +1 ) = p + 1. This proves that for the case n=0 the formula (4.161) gives the seed (4.163). Now, consider the formula (4.161) for the cases of n=1, 2, 3, ..., р: 0 0 0 Lp (1) = x1 + x 2 + ... + x p +1 ( ) L ( 3 ) = ( x ) + ( x ) + ... + ( x ) Lp ( 2 ) = ( x1 ) + ( x 2 ) + ... + x p +1 2 2 3 p 1 3 2 3 p +1 2 ... p p p Lp ( p ) = ( x1 ) + ( x 2 ) + ... + x p +1 . ( (4.165) ) Then, according to Theorem 4.2, all expressions (4.165) are equal to 1. This proves that the expression (4.161) is valid for the cases n=1, 2, 3, …, p. To prove the validity of the recursive relation (4.162) for the general case of n, we can use the identity (4.76) and represent the formula (4.161) as follows: Alexey Stakhov THE MATHEMATICS OF HARMONY 222 ( ) n −1 n −1 n −1 Lp ( n ) = ( x1 ) + ( x 2 ) + ... + x p +1 n − p −1 n − p −1 n − p −1 . + ( x2 ) + ... + x p +1 + ( x1 ) ( (4.166) ) Using the definition (4.161), we can rewrite the expression (4.166) in the form of the recursive relation (4.162). Our reasoning resulted in the discovery of new recursive numerical se quences Lucas pnumbers given by the recursive relation (4.162) at the seeds (4.163) and (4.164). It is clear that for the case p=1 this recursive relation is reduced to the recursive relation for the classical Lucas numbers. Let us study the partial cases of the Lucas pnumbers for the cases p=2, 3, 4. 4.10.2. Binet Formula for the Lucas 2Numbers For the case p=2, the formula (4.161) can be presented in the form below: L2 ( n ) = ( x1 ) + ( x2 ) + ( x3 ) . n n n (4.167) This formula defines the Lucas 2numbers L2(n). If we substitute the expressions for the roots x1, x2, x3, given by (4.107) (4.110) into (4.167) we can rewrite the formula (4.167) as follows: n 1 1 i 3 h 2 h 2 1 h L2 ( n ) = + + + − − + + − 2 6 3h 6 3h 3 12 3h 3 n h 1 1 i 3 h 2 + − − + − − . 2 6 3h 12 3h 3 n (4.168) For the case p=2, the recursive relation (4.162) and the seeds (4.163) and (4.164) are reduced to the following: L2(n)=L2(n1)+L2(n3) (4.169) L2(0)=3 (4.170) L2(1)=L2(2)=1. (4.171) Then, using the recursive relation (4.169) at the seeds (4.163) and (4.164), we can calculate all elements of the Lucas 2series: 3, 1, 1, 4, 5, 6, 10, 15, 21, 31, 46, 67, 98, 144, … . 4.10.3. Binet Formula for the Lucas 3Numbers For the case p=3, the formula (4.161) takes the following form: (4.172) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 223 (4.173) L3 ( n ) = ( x1 ) + ( x2 ) + ( x3 ) + ( x4 ) . This formula is an analytical expression for the Lucas 3numbers L3(n). Using the numerical values for the roots x1, x2, x3, x4 given by (4.142), we can represent the formula (4.173) in the following numerical form: n n n n L3(n)=(1.380)n+(0.819)n+(0.219+0.914i)n+(0.2190.914i)n. (4.174) For the case p=3 the recursive relation (4.162) and the seeds (4.163) and (4.164) are reduced to the following: L3(n)= L3(n1)+ L3(n4) (4.175) L3(0)=4 (4.176) (4.177) L3(1)= L3(2)=L3(3)=1. The recursive relation (4.175) at the seeds (4.176) and (4.177) generates the following Lucas 3series: 4,1,1,1,5,6,7,8,13,19,26,34,47,66,… . (4.178) 4.10.4. Binet Formula for the Lucas 4Numbers For the case p=4, the formula (4.161) takes the following form: n n n n n L4 ( n ) = ( x1 ) + ( x2 ) + ( x3 ) + ( x4 ) + ( x5 ) . (4.179) This formula sets the Lucas 4numbers L4(n) in analytical form. Substituting the analytical representations of the roots x1, x2, x3, x4, x5 giv en by (4.150) into (4.179), we obtain the following formula: h L4 ( n ) = 6 n + n 2 1 i 3 1 i 3 + − + + 2 2 2 h 2 n n n (4.180) h 1 i 3 h 2 h 1 i 3 h 2 − + − − + − + − − − . 2 6 6 12 6 2 6 6 12 6 For the case p=4, the recursive relation (4.162) and the seeds (4.163) and (4.164) are reduced to the following: L4(n)= L4(n1)+ L4(n5) (4.181) L4(0)=5 (4.182) L4(1)= L4(2)=L4(3)=L4(4)=1. (4.183) The recursive relation (4.181) at the seeds (4.182) and (4.183) generates the following Lucas 4series: 5, 1, 1, 1, 1, 6, 7, 8, 9, 10, 16, 23, 31, 40, … . (4.184) THE MATHEMATICS OF HARMONY Alexey Stakhov 224 4.10.5. The “Extended” Lucas рNumbers Above we introduced the socalled “extended” Fibonacci pnumbers that are given for the negative values of n (see Table 4.2). By analogy, we can in troduce the “extended” Lucas pnumbers, if we extend the Lucas pnumbers to the side of the negative values of n. With this purpose in mind, we shall find some general properties of such “extended” sequences. For the calculation of the Lucas pnumbers Lp(0), Lp(1), Lp(2),…, Lp(p),…, Lp(2p+1), …, which correspond to the nonnegative values of n=1,2,3,…, we represent the re cursive relation (4.162) as follows: Lp(np1)= Lp(n)Lp(n1). In particular, for the case n=p the formula (4.185) is (4.185) (4.186) Lp(1)= Lp(p)Lp(p1). Now, consider the formula (4.185) for the different values of p. For p=1, the formulas (4.185) and (4.186) take the following forms, respectively: L1(n2)=L1(n)L1(n1). (4.187) L1(1)=L1(1)L1(0). As L1(1)=1 and L1(0)=2, it comes from (4.188) that (4.188) L1(1)=1. (4.189) Using the recursive relation (4.186), we can calculate all values of the Lucas numbers L1(n) for the negative values of n=1, 2, 3, … and then repre sent the classical Lucas numbers L1(n) as is shown in Table 4.3. Table 4.3. The “extended” classical Lucas numbers n 5 4 3 2 1 0 1 2 3 4 5 L1(n) 11 7 4 3 1 2 1 3 4 7 11 For the case p=2, the formulas (4.185) and (4.186) take the following form: L2(n3)=L2(n)L2(n1) (4.190) L2(1)=L2(2)L2(1). Taking into consideration (4.171), we can write: (4.191) (4.192) L2(1)=0. Using (4.170) and calculating numerical values of the Lucas 2numbers L2(n) for the nonnegative n=0, 1, 2, 3, 4, ... , we can get the following numerical sequence: L2(n) (n≤0): 3, 0, 2, 3, 2, 5, 1, 7, 6, 6, … . (4.193) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 225 For the case p=3, the formulas (4.185) and (4.186) take the following form: L3(n4)=L3(n)L3(n1) (4.194) L3(1)=L3(3)L3(2). (4.195) Using (4.170) and (4.194) and (4.195), we can calculate numerical values of the Lucas 3numbers L3(n) for the nonnegative n=0, 1, 2, 3, 4, ..., we can get the following numerical sequence L3(n) (n≤0): 4, 0, 0, 3, 4, 0, 3, 7, 4, 3, … . Of course, by using the general recursive relation (4.185), we can calcu late the rest of the values of Lp(n) for the nonnegative values of n. In Table 4.4 we can see the “extended” Lucas numbers Lp(n) for the cases p=1, 2, 3, 4. Table 4.4. The “extended” Lucas pnumbers n 6 5 4 3 2 1 0 1 2 3 4 5 6 L1(n) 18 11 7 4 3 1 2 1 3 4 7 11 18 L2(n) 10 6 5 4 1 1 3 0 2 3 2 5 1 L3(n) 7 6 5 1 1 1 4 0 0 3 4 0 3 L4(n) 7 6 1 1 1 1 5 0 0 0 4 5 0 4.10.6. Identities for the Sums of the Lucas рNumbers Once again, consider the recursive relation (4.169). Decomposing the Lucas 2number L2(n1) in (4.169) according to the same recursive relation (4.169), that is, representing L2(n1) in the form L2(n1)=L2(n2)+L2(n4), we can represent the formula (4.169) in the following form: (4.196) L2(n)=L2(n2)+ L2(n3)+L2(n4). This means that the sum of the three sequential Lucas 2numbers is al ways equal to the Lucas 2number that is two positions from the senior Lucas 2number of the sum. If we use a similar approach for the Lucas рnumbers in the general case, we obtain the following general identity: Lp(n)=Lp(np)+Lp(np1)+Lp(np2)+…+Lp(n2p). (4.197) Note that the identity (4.197) is valid for all “extended” Lucas pnumbers. Now, let us examine the sum of the first n Lucas pnumbers: Alexey Stakhov THE MATHEMATICS OF HARMONY 226 Lp(1)+Lp(2)+Lp(3)+…+Lp(n). (4.198) In order to get the required result, we write down the basic recursive rela tion (4.162) for the Lucas pnumbers in the following form: Lp(n)=Lp(n+p+1)Lp(n+p). Using (4.199), we can write the following equalities: (4.199) Lp (1) = Lp ( 2 + p ) − Lp (1 + p ) Lp ( 2 ) = Lp ( 3 + p ) − Lp ( 2 + p ) Lp ( 3 ) = Lp ( 4 + p ) − Lp ( 3 + p ) ... Lp ( n − 1) = Lp ( n + p ) − Lp ( n + p − 1) Lp ( n ) = Lp ( n + p + 1) − Lp ( n + p ) . Summing term by term the lefthand and righthand parts of these equal ities, we obtain the following formula: Lp(1)+Lp(2)+Lp(3)+…+Lp(n)= Lp(n+p+1)Lp(1+p). (4.200) It follows from the recursive relation (4.162) and the seeds (4.163) and (4.164) that (4.201) Lp(1+p)=Lp(p)+Lp(0)=1+p+1=p+2. The following formula for the sum (4.200) follows from (4.201): Lp(1)+Lp(2)+Lp(3)+…+Lp(n)= Lp(n+p+1) p2. (4.202) 4.10.7. The Ratio of Adjacent Lucas pNumbers Above we found that the Fibonacci pnumbers are closely connected with the golden pproportion. In particular, the limit of the ratio Fp(n)/Fp(n1) aims for the golden pproportion. There is a question: what is the limit of the ratios of adjacent Lucas pnumbers? Introduce the following definition: L p ( n) lim = x. (4.203) n →∞ L (n − 1) p Using the recursive relation (4.162), we can represent the ratio of the ad jacent Lucas pnumbers as follows: L p ( n) Lp (n − 1) = 1+ 1 Lp (n − 1) Lp (n − p − 1)) = Lp (n − 1) + Lp (n − p − 1) = 1+ Lp (n − p − 1) 1 Lp (n − 1) ⋅ Lp (n − 2) ⋅⋅⋅ Lp (n − p) Lp (n − 2) ⋅ Lp (n − 3) " Lp (n − p − 1) . (4.204) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 227 Taking into consideration the definition (4.203), for the case n→∞ we can exchange the formula (4.204) by the algebraic equation (4.42) with the positive root τp the golden pproportion. It follows from this reasoning that lim L p ( n) n →∞ L p (n − 1) = τp. (4.205) 4.11. The “Metallic Means Family” by Vera W. de Spinadel 4.11.1. The Generalized Fibonacci Sequences Recently the Fibonacci numbers and the golden mean were generalized by the Argentinean mathematician Vera W. de Spinadel who is the author of the original book in this regard [42]. Spinadel generalized the Fibonacci re cursive relation F(n+1)=F(n)+F(n1) as follows: G(n+1)=pG(n)+qG(n1), (4.206) where p and q are natural numbers. Consider the examples of the generalized Fibonacci sequences of the kind (4.206). If we assume that p=2 and q=1 in (4.206) and begin from the seeds G(1)=G(2)=1, then the recursive relation (4.206) generates the following generalized Fibonacci numbers: 1, 1, 3, 7, 17, 41, 99, 140, … . (4.207) For the case p=3 and q=1, and G(1)=G(2)=1 the generalized Fibonacci numbers are: 1, 1, 1 , 4, 13, 43, 142, 469, … . (4.208) We can represent the recursive relation (4.206) in the following form: G(n + 1) G(n − 1) q = p+q = p+ . G ( n) G ( n) G ( n) (4.209) G(n − 1) If we denote by x the limit of the ratio of two adjacent generalized Fi bonacci numbers, that is, G(n + 1) x = lim , n →∞ G(n) then we can represent the expression (4.209) as follows: Alexey Stakhov THE MATHEMATICS OF HARMONY 228 x = p+ q x (4.210) or x2pxq=0. This algebraic equation has the following positive solution: x= p + p 2 + 4q . 2 This means that G(n + 1) p + p2 + 4q . = n →∞ G(n) 2 lim (4.211) (4.212) (4.213) 4.11.2. The Metallic Means Family Above we have introduced the quadratic algebraic equation (4.211). Vera W. de Spinadel proved that Eq. (4.211) gives an infinite set of positive qua dratic irrationals for the different values of p and q. They are all given by the formula (4.212) and together make the Metallic Means Family (MMF). Let us denote any member of the MMF by σqp , where p and q take their values from the set of natural numbers, that is, p=1, 2, 3, … ; q=1, 2, 3, … . Consider special cases of Eq. (4.211). We start from the case q=1, that is, from the following algebraic equation: (4.214) x2px1=0. It is convenient to represent the members of the MMF in the form of a continued fraction. It is clear that for the case p=1 Eq. (4.214) is reduced to the simplest algebraic equation x2x1=0 with the positive root equal to the golden mean. Equation (4.215) can be written in the form: (4.215) 1 . (4.216) x By substituting the golden mean τ for x in Eq. (4.216), we obtain the following representation of τ : 1 τ =1+ . (4.217) τ It comes from (4.217) that the following representation of the golden x =1+ Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 229 mean in the form of a continued fraction: 1 . τ =1+ 1 1+ (4.218) 1 1+ 1 + ... Note that the representation (4.218) was considered in Chapter 1. In mathematics, for the compact representation of the continued fractions the following representation is used: ϕ = [1,1,1,...] = 1 . (4.219) Now, assume that p=2 in Eq. (4.214), that is, consider the following equation: (4.220) x22x1=0. Then, according to the above definition we can denote the positive root of Eq. (4.220) by σ12 . We can represent Eq. (4.220) in the form: 1 x =2+ . (4.221) x If we substitute 2+(1/x) for x on the righthand part of (4.221), we obtain the following representation of x: x =2+ 1 . 1 (4.222) x Continuing this process ad infinitum, we obtain the representation of σ12 2+ in the form of the following continued fraction: 1 . σ12 = 2 + 1 2+ (4.223) 1 2+ 2 + ... The positive quadratic irrational number σ12 given by (4.223) was named by Spinadel the Silver Mean. By analogy with the golden mean (4.219), the Silver Mean (4.223) can be represented in the following compact form: σ12 = [ 2, 2, 2,...] = 2 . (4.224) Using the formula (4.212), we can write the analytical representation of the silver mean σ12 as follows: σ12 = 1 + 2 = 2 . (4.225) If we assume p=3, then Eq. (4.214) takes the following form: x 3x1=0. 2 (4.226) Alexey Stakhov THE MATHEMATICS OF HARMONY 230 The positive root of this equation is called the Bronze Mean σ13 ; it can be represented in the form of the following continued fraction: 1 σ13 = 3 + 1 3+ (4.227) 1 3+ 3 + ... or σ12 = [ 3, 3, 3,...] = 3 . (4.228) Using the formula (4.212), we can write the analytical representation of the bronze mean σ13 as follows: 3 + 13 = 3 . (4.229) 2 For the cases where p=4, 5, 6, 7, 8, 9, 10, respectively, we can find the following analytical representations of the corresponding Metallic Means: σ13 = 5 + 29 7 + 53 ; σ16 = 3 + 10 ; σ17 = ; 2 2 9 + 85 σ18 = 4 + 17 ; σ19 = ; σ110 = 5 + 26 . 2 It is clear that all Metallic Means of the kind σ1p have the following gen σ14 = 2 + 5 ; σ15 = eral representation in the form of the periodic continued fraction: 1 = [ p ]. σ1p = p + 1 p+ 1 p+ p + ... (4.230) 4.11.3. Other Types of the Metallic Means If we assume that p=1 in Eq. (4.211), then we obtain the following equation: (4.231) x2xq=0, where q is a natural number. The positive roots of this equation generate a new class of the Metallic Means denoted by σ1q . Note that for the case q=1, Eq. (4.231) is reduced to the golden equation (4.215). For the case q=2, Eq. (4.231) takes the following form: x2x2=0. (4.232) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 231 The number σ1q = 2 is a positive root of this equation. Spinadel calls this number the Copper Mean. Using a traditional representation of the continued fraction, we can represent the Copper Mean as follows: σ1q = 2 = 2, 0 . If we assume that q=3 in Eq. (4.231), then we obtain the following equation: 2 x x3=0. This equation results in the Nickel Mean (4.233) 1 + 13 = 2, 3 . (4.234) 2 By analogy, we can find the next representations of the other Metallic Means of the type σ1q . σ13 = 4.11.4. PisotVijayaraghavan Numbers and Metallic Means Vera W. de Spinadel paid attention to the fact that her Metallic Means have a direct relation to PisotVijayaraghavan numbers or PVnumbers [152]. It is well known that the PVnumbers are positive roots of the following alge braic equation: (4.235) x m = am −1 x m −1 + ... + ai x i + ... + a1 x + a0 , where ai are integers. The golden mean τ = 1 + 5 2 is the example of the PVnumbers be cause the golden mean is the positive root of Eq. (4.215), which is a partial case of (4.235). Also, all the golden pproportions tp are PVnumbers because Eq. (4.42) is a partial case of (4.235). The number Q1=1.324… a positive root of the equation x3x1=0 is also a PVnumber. This number is called a Plastic Constant. It is proved that Eq. (4.215) with the root τ = 1 + 5 2 appears in qua sicrystal structures. In addition, the following algebraic equations which are partial cases of (4.235) appear in quasicrystal structures: ( ) ( ) x 2 − 2x − 1 = 0 → γ = 1 + 2 (4.236) (4.237) x 2 − 4x + 1 = 0 → δ = 1 + 3. τ = 1 + 5 2 (the golden mean) Note that the particular PVnumber corresponds to pentagonal and decagonal quasilattices, while another PV number γ = 1 + 2 (the Silver Mean) corresponds to the octagonal quasilat tice, and the PVnumber δ = 1 + 3 corresponds to the case of the dodecago nal quasilattice. ( ) Alexey Stakhov THE MATHEMATICS OF HARMONY 232 The above examples demonstrate that Spinadel’s Metallic Means are of great theoretical importance for algebra and crystallography. 4.11.5. “Silver Means” by Jay Kappraff The American researcher Jay Kappraff – the author of the interesting books [47, 50 ] – developed an approach to the generalization of Fibonacci numbers and the golden mean similar to Spinadel’s Metallic Means. He considered the recursive relation G(n+1)=N G(n)+G(n1) (4.238) that is a partial case of the recursive relation (4.206) for the case q=1 and p=N. This recursive relation gives different generalized Fibonacci numbers at different seeds. From the recursive relation (4.238) Kappraff derived the fol lowing algebraic equation 1 = N, (4.239) x which is the other form of the equation (4.214). It is clear that for the case N=1 and the seeds G(0)=0 and G(1)=1 the recursive relation (4.238) generates the classical Fibonacci numbers 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, ... . The ratios of the adjacent numbers in this series converge to the golden mean τ a positive root of the golden equation x2x1=0. Kap praff names the classical golden mean τ the 1st Silver Mean of Type 1 and denotes it by SM1(1). For the case N=2, the recursive relation (4.238) at the seeds G(1)=1 and G(2)=2 generates Pell’s Sequence 1,2,5,12,29,... . The ratios of the ad jacent numbers in Pell’s sequence converge to the positive root of the equation x 22x1=0 Kappraff names the positive root of this equation θ = 1 + 2 = 2.414... the 2nd Silver Mean of Type 1 and denotes it by SM1(2). x− 4.12. Gazale Formulas 4.12.1. The Generalized Fibonacci mNumbers Independently of Spinadel and Kappraff, the idea of generalized Fibonac ci numbers was developed in the book [45] written by Egyptian mathemati cian and engineer Midchat Gazale. Gazale considers the recursive relation Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 233 Fm(n+2)=m Fm(n+1)+Fm(n), (4.240) where m is a positive real number and n=0,±1,±2,±3,… . Note that the recur sive formula (4.240) is similar to the recursive formula (4.206) used by Spi nadel, when we take p=m and q=1 in (4.206). However, in contrast to (4.206), where the coefficients p and q are integers, the coefficient m in the recursive formula (4.240) used by Gazale is a positive real number m>0. We will name a positive real number m, used in the recursive relation (4.240), an order of the recursive relation (4.240). If we take the seeds (4.241) Fm(0)=0, Fm(1)=1 and then use the recursive relation (4.240) for a given m>0, we obtain an infinite number of the recursive numerical sequences called Generalized Fi bonacci Numbers of Order m or simply Fibonacci mnumbers. Note that for the case m=1 the recursive relation (4.240) and the seeds (2.41) can be represented, respectively, as follows: F1(n+2)=F1(n+1)+F1(n) (4.242) F1(0)=0, F1(1)=1. (4.243) It is clear that the recursive relation (4.242) with the seeds (4.243) gener ates the classical Fibonacci numbers. For the case of m=2 the recursive formula (4.240) and the seeds (4.241) are reduced to the following: F2(n+2)=2F2(n+1)+F2(n) (4.244) (4.245) F2(0)=0, F2(1)=1. This case generates the socalled Pell numbers: 0, 1, 2, 5, 12, 29, … . If we take m = 2, then the recursive relation (4.240) at the seeds (4.241) generates the following recursive numerical sequence: 0, 1, 2 , 3, 4 2 , 11, 15 2 , 41, 56 2 ,.... 4.12.2. The Generalized Golden Mean of Order m Let us represent the recursive relation (4.240) as follows: Fm (n + 2) 1 =m+ . Fm (n + 1) Fm (n + 1) (4.246) Fm (n) For the case n→∞, the expression (4.246) is reduced to the following qua dratic equation: (4.247) x2mx1=0. Alexey Stakhov THE MATHEMATICS OF HARMONY 234 Equation (4.247) has two roots a positive root 4 + m2 + m 2 and a negative root x1 = (4.248) − 4 + m2 + m . 2 If we sum up (4.248) and (4.249), we obtain: x2 = (4.249) (4.250) x1+ x2=m. If we substitute the root (4.248) for x in Eq. (4.247), we obtain the follow ing identity: (4.251) x12 = mx1 + 1. If we multiply or divide repeatedly all terms of the identity (4.251) by x1, we obtain the following general identity: (4.252) x1n = mx1n −1 + x1n −2 , where n=0,±1,±2,±3,… . Using similar reasoning for the root x2, we obtain the following identity for the root x2: (4.253) x2n = mx2n −1 + x2n −2 . Denote a positive root x1 by Fm and name it the Golden Mean of Order m or the Golden mProportion. The golden mproportion Fm has the following analytical expression: 4 + m2 + m (4.254) . 2 Note that for the case m=1, the golden mproportion coincides with the classical golden mean Φ1 = 1 + 5 2 . Let us express the root x2 by the golden mproportion Φm. After simple transformation of (4.249) we can write the root x2 as follows: Φm = ( x2 = ) − 4 + m2 + m −4 1 = =− . 2 2 Φm 2( 4 + m + m) (4.255) Substituting Fm for x1 and (1/Fm) for x2 in (4.250), we obtain: m = Φm − 1 / Φm , where Φm is given by (4.254) and 1/Φm is given by the formula: (4.256) 1 4 + m2 − m = . 2 Φm (4.257) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 235 Using the formulas (4.254) and (4.257), we can write the following identity: 1 Φm + = 4 + m2 . (4.258) Φm It is also easy to prove the following identity: Φmn = mΦmn −1 + Φmn −2 , where n=0, ±1, ±2, ±3,… . (4.259) 4.12.3. Two Surprising Representations of the Golden mProportion For the case n=2, the identity (4.259) can be represented in the form: (4.260) Φm2 = 1 + mΦm . The following representation of the golden mproportion Φm comes from (4.260): Φm = 1 + mΦm . (4.261) Substituting 1+ mΦm for Φm on the righthand part of (4.261), we obtain: Φm = 1 + m 1 + Φm . (4.262) Continuing this process ad infinitum, that is, substituting repeatedly 1+ mΦm for Φm on the righthand part of (4.262), we obtain the following sur prising representation of Φm: Φm = 1 + m 1 + m 1 + m ... . Now, represent the identity (4.260) in the form: 1 Φm = m + . Φm (4.263) (4.264) Substituting m+(1/Φm) for Φm on the righthand part of (4.264), we obtain: 1 Φm = m + . 1 (4.265) m+ Φm Continuing this process ad infinitum we obtain the following surprising representation of the golden mproportion: 1 Φm = m + . 1 m+ (4.266) 1 m+ m + ... Alexey Stakhov THE MATHEMATICS OF HARMONY 236 Note that for the case of m=1, the representations (4.263) and (4.28) coin cide with the wellknown representations of the classical golden mean in the forms (1.20) and (1.24), respectively. 4.12.4. A Derivation of the Gazale Formula The formula (4.240) at the seeds (4.241) defines the Fibonacci mnumbers Fm(n) by recursion. We can represent the numbers Fm(n) in analytical form us ing the golden mproportion Φm. Let us represent the Fibonacci number Fm(n) by the roots x1 and x2 in the form: (4.267) Fm (n) = k1 x1n + k2 x2n , where k1 and k2 are constant coefficients that are the solutions to the follow ing system of algebraic equations: Fm (0) = k1 x10 + k2 x20 = k1 + k2 . (4.268) 1 1 Fm (1) = k1 x1 + k2 x2 = k1Φm − k2 (1 / Φm ) Taking into consideration that Fm(0)=0 and Fm(1)=1, we can rewrite the system (4.268) as follows: k1 = k2 (4.269) k1Φm + k1 (1 / Φm ) = k1 ( Φm + 1 / Φm ) = 1. (4.270) Taking into consideration (4.269) and (4.270) and also the identity (4.258), we can find the following formulas for the coefficients k1 and k2: 1 1 ; k2 = − . k1 = (4.271) 4 + m2 4 + m2 Taking into consideration (4.271), we can write the formula (4.267) as follows: 1 1 1 Fm (n) = x1n − x2n = x1n − x2n . (4.272) 2 2 2 4+m 4+m 4+m Taking into consideration that x1=Φm and x2=1/Φm, we can rewrite the formula (4.272) as follows: Φ n − (−1 / Φm )n Fm (n) = m (4.273) 4 + m2 or ( ) n n 2 m − 4 + m2 m + + m 4 − . Fm (n) = 2 2 2 4+m 1 (4.274) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 237 For the partial case m=1, formula (4.274) is reduced to the formula: n n 1 1 + 5 1 − 5 F1 (n) = − (4.275) 5 2 2 called the Binet formula. This formula was obtained by Binet in 1843, although the result was known to Euler, Daniel Bernoulli, and de Moivre more than one century earlier. In particular, de Moivre obtained this formula in 1718. For the case m=2, the formula (4.274) takes the following form: n n 1 F2 (n) = 1+ 2 − 1− 2 . (4.276) 2 2 ( ) ( ) Note that this formula was obtained for the first time by the English math ematician John Pell (16101685). For the cases m=3 and m = 2 , the formula (4.274) takes the following forms, respectively: n n 1 3 + 13 3 − 13 F3 (n) = − (4.277) 2 2 13 n n 1 2 + 6 2 − 6 F 2 (n) = . − (4.278) 2 2 6 Thus, the Egyptian mathematician Midhat J. Gazale obtained [45] the unique mathematical formula (4.274), which includes as partial cases the Bi net formula for Fibonacci numbers (4.275) for the case m=1 and Pell’s formu la (4.276) for the case m=2. This formula generates an infinite number of the Fibonacci mnumbers because every positive real number m generates its own Generalized Binet Formula (4.274). Taking into consideration the uniqueness of the formulas (4.272)(4.274) we will name this formula the Gazale Formula for the Fibonacci mnumbers. 4.13. Fibonacci and Lucas mNumbers 4.13.1. Fibonacci mNumbers Let us prove that Gazale formulas (4.272)(4.274) really express all Fi bonacci mnumbers given by the recursive formula (4.240) at the seeds (4.241). In fact, for the case n=0 the formula (4.274) gives the Fibonacci mnumber Alexey Stakhov THE MATHEMATICS OF HARMONY 238 Fm(0)=0 that corresponds to the seeds (4.241). For the case n=1, we can rewrite the formula (2.274) as follows: m + 4 + m2 m − 4 + m2 =1 − 2 2 4 + m2 1 Fm (1) = that corresponds to the seeds (4.241). This means that the formula (4.276) does express the seeds (4.241). Suppose that the formula (4.274) is valid for a given n (the inductive hy pothesis) and prove that this formula is valid for the case n+1, that is, 1 Fm (n + 1) = x1n +1 − x2n +1 . (4.279) 2 4+m Using the identities (4.252) and (4.253), we can represent the formula (4.279) as follows: m 1 Fm (n + 1) = x1n − x2n ) + ( ( x1n−1 − x2n−1 ) 2 2 4+m 4+m (4.280) = mFm ( n ) + Fm ( n − 1) . ( ) Thus, the formula (4.280), in fact, defines the Fibonacci mnumbers given by the recursive relation (4.240) at the seeds (4.241). Note that the formula (4.274) defines all Fibonacci mnumbers in the range n=0, ±1, ±2, ±3, … . Let us find some surprising properties of the Fibonacci m numbers. First of all, compare Fm(n) and Fm(n). We can write the formula (4.273) as follows: Fm (n) = Φmn − (−1)n Φm− n . (4.281) 4 + m2 Let us consider the formula (4.281) for the negative values of n, that is, Fm (− n) = Φm− n − (−1)− n Φmn . (4.282) 4 + m2 Comparing the expression (4.281) and (4.282) for the even (n=2k) and odd (n=2k+1) values of n, we find: (4.283) Fm(2k)= Fm(2k) and Fm(2k+1)=Fm(2k1). This means that the sequences of the Fibonacci mnumbers in the range n=0, ±1, ±2, ±3, … are symmetrical sequences with respect to the Fibonacci mnumber Fm(0)=0 except that the Fibonacci mnumbers Fm(2k) and Fm(2k) are opposite in sign. In Table 4.5 we can see the Fibonacci mnumbers for the cases m=1, 2, 3, 4. Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 239 Table 4.5. Fibonacci mnumbers (m=1, 2, 3, 4) m 4 3 2 1 0 1 2 3 4 1 3 2 1 1 0 1 1 2 3 2 12 5 2 1 0 1 2 5 12 3 33 10 3 1 0 1 3 10 33 4 72 17 4 1 0 1 4 17 72 4.13.2. The Generalized Cassini Formula Next let’s find a fundamental formula that connects three adjacent Fi bonacci mnumbers. For the case m=1, this formula is known as the Cassini Formula. Remember that this formula for the classical Fibonacci numbers F1(n) has the following form: F12 ( n ) − F1 ( n − 1) × F1 ( n + 1) = ( −1) n +1 . (4.284) It is easy to prove the following general identity for the Fibonacci mnumbers: n +1 Fm2 ( n ) − Fm ( n − 1) × Fm ( n + 1) = ( −1) . (4.285) For example, for the case m=2 the Fibonacci mnumbers F2(5)=29, F2(4)=12 and F2(3)=5 are connected by the following correlation: (12)2 (29×5)=1 and for the case m=3 the Fibonacci mnumbers F3(4)=33, F2(3)=10 and F3(3)=3 are connected by the following correlation: (10)2(33×3)=1. Note that both examples correspond to the general formula (4.285). 4.13.3. Lucas mNumbers Once again, consider the formula (4.267) that defines the Fibonacci mnum bers. By analogy with the classical Lucas numbers we can consider the formula (4.286) Lm ( n ) = x1n + x2n . It is clear that for the case m=1, this formula defines the classical Lucas numbers: 2, 1, 3, 4, 7, 11, 18, … . Let us assume that this formula defines the Generalized Lucas Numbers of Order m or simply Lucas mNumbers. For a giv en m we can find some peculiarities of the Lucas mnumbers. First of all, cal culate the seeds of the Lucas mnumbers given by (4.286). In fact, for the cases n=0 and n=1 we have from (4.286), respectively: Lm ( 0 ) = x10 + x20 = 1 + 1 = 2 (4.287) Lm (1) = x11 + x21 = Φm + ( −1 / Φm ) = m. (4.288) THE MATHEMATICS OF HARMONY Alexey Stakhov 240 Note that for the case m=1, the seeds (4.287) and (4.288) come for the seeds of the classical Lucas numbers: L1(0)=2, L1(1)=1. Using (4.252) and (4.253), we can represent the formula (4.286) as follows: Lm ( n ) = x1n + x2n = mx1n −1 + x1n −1 + mx2n −1 + x2n −1 (4.289) = m x1n −1 + x2n −1 + x1n −1 + x2n −1 . ( ) ( ( ) ) ( ) Taking into consideration the definition (4.286), we can rewrite (4.289) in the following form of a recursive relation: (4.290) Lm(n)=mLm(n1)+Lm(n2). It is clear that the recursive relation (4.290) at the seeds (4.287) and (4.288) gives the Lucas mnumbers in recursive form. If we substitute x1=Φm and x2=1/Φm in the formula (4.286), we can repre sent the Lucas mnumbers in the following analytical form: n Lm ( n ) = Φmn + ( −1 / Φm ) . (4.291) Although this formula is absent in Gazale’s book [45], we will name this important formula the Gazale Formula for Lucas mNumbers. We can rewrite the formula (4.291) as follows: n Lm ( n ) = Φmn + ( −1) Φm− n . (4.292) Let us write the formula (4.292) for the negative values of n, that is, n Lm ( − n ) = Φm− n + ( −1) Φmn . (4.293) Comparing the expressions (4.292) and (4.293) for even (n=2k) and odd (n=2k+1) values of n, we obtain: (4.294) Lm(2k)=Lm(2k) and Lm(2k+1)=Lm(2k1). This means that the sequences of Lucas mnumbers in the range n=0, ±1, ±2, ±3,… are symmetrical sequences with respect to the Lucas mnumber Lm(0)=2 except that the Lucas mnumbers Lm(2k+1) and Lm(2k1) are oppo site by sign. In Table 4.6 we can see the Lucas mnumbers for the cases m=1, 2, 3, 4. Table 4.6. The Lucas mnumbers (m=1, 2, 3, 4) m 4 3 2 1 0 1 2 3 4 1 7 4 3 1 2 1 3 4 7 2 34 14 6 2 2 2 6 14 34 3 119 36 11 3 2 3 11 36 119 4 322 76 18 4 2 4 18 76 322 Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 241 Note that for the case m=1 the Lucas mnumbers coincide with the classical Lucas numbers and for the case m=2 coincides with the PellLucas numbers. Note similar results obtained by Kappraff and Adamson in [153]. 4.14. On the mExtension of the Fibonacci and Lucas pNumbers 4.14.1. A Recursive Relation for the Fibonacci (p,m)Numbers Now, we define the recursive relation for the mextension of the Fibonac ci pnumbers [154]. For a given integer p>0 and positive real number m>0 the recursive relation is given as follows Fp,m(n)=mFp,m(n1)+Fp,m(np1) with initial conditions (4.295) Fp,m(0)=a0, Fp,m(1)=a1, Fp,m(2)=a2, …, Fp,m(p)=ap, where a0, a1, a2, … , ap are integer, real or complex numbers. In particular, we can take these initial conditions as follows (4.296) Fp,m(0)=0, Fp,m(k)=mk1, where k=1,2,3,…,p. We name a new class of recursive numerical sequences given by (4.295) at the seeds (4.296) an mextension of the Fibonacci pnumbers or simply Fibonacci (p, m)numbers. It is clear that the recursive formula (4.295) at the seeds (4.296) defines a more general class of recursive numerical sequences than the Fibonacci pnum bers or the Fibonacci mnumbers. Note that for the case m=1 the Fibonacci (p,m)numbers coincide with the Fibonacci pnumbers, that is, Fp,1(n)= Fp (n), and for the case p=1, the Fibonacci (p,m)numbers coincide with the Fibonacci mnumbers, that is, F1,m(n)=Fm (n). For the cases p=1 and m=1, the Fibonacci (p, m)numbers coincide with the classical Fibonacci numbers. 4.14.2. Some Properties of the Fibonacci (p,m)Numbers Let us consider the mextension of the Fibonacci pnumbers given by (4.295) with the initial conditions of (4.296). Let us calculate the set of Fi bonacci (p,m)numbers for the negative values of the argument m Fp,m(1), Fp,m(2), … , Fp,m(p), … , Fp,m(2p+1). THE MATHEMATICS OF HARMONY Alexey Stakhov 242 According to (4.296), we have: F p,m(p+1)=mp and Fp,m(p)=mp1, thus Fp,m(0)=0. Continuing in this way, we obtain (4.297) Fp,m(1)=Fp,m(2)=…=Fp,m(p+1)=0. Let us write the mextension of the Fibonacci pnumber Fp,m(1) in the form Fp,m(1)=mFp,m(0)+Fp,m(p). Then, we obtain Fp,m(p)=1. Using (4.295), (4.296) and (4.298), we have (4.298) Fp,m(p1)= Fp,m(p2)=…= Fp,m(2p+1)= 0. Also, we have (4.299) (4.300) Fp,m(2p)=m, Fp,m(2p1)=1, Fp,m(2p2)= 0,…. For the case m=2 we obtain the 2extension of Fibonacci pnumbers called Pell pnumbers. Table 4.7 gives the values of Pell pnumbers for the cases of p=1, 2, 3, 4. Table 4.7. Pell pnumbers (m=2; p=1,2,3,4) n 5 4 3 2 1 0 1 2 3 4 5 F1,2(n) 29 12 5 2 1 0 1 2 5 12 29 F2,2(n) 22 9 4 2 1 0 0 1 0 2 1 F3,2(n) 17 8 4 2 1 0 0 0 1 0 0 F4,2(n) 16 8 4 2 1 0 0 0 0 1 0 4.14.3. Characteristic Algebraic Equation for the Fibonacci (p, m) Numbers Let us represent the recursive relation (4.295) in the form: Fp ,m (n) Fp ,m (n −1) = m+ = 1 Fp ,m (n −1) mFp ,m (n −1) + Fp ,m (n − p −1) = m+ Fp ,m (n − p −1) Fp ,m (n −1) 1 Fp ,m (n −1) ⋅ Fp ,m (n − 2) ⋅⋅⋅ Fp ,m (n − p) . (4.301) Fp ,m (n − 2) ⋅ Fp ,m (n − 3) " Fp ,m (n − p −1) Suppose that lim n→∞ F Fp,m (n) p ,m (n − 1) = x. (4.302) Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 243 Taking into consideration the definition (4.302) and directing n → ∞, we can replace the expression (4.301) with the following algebraic equation 1 x =m+ p , x whence it appears xp+1mxp1=0. (4.303) We will name the equation (4.303) a Characteristic Equation for the Fi bonacci (p, m)numbers. Note that for the case m=1, Eq. (4.303) is reduced to Eq. (4.42) – the characteristic equation for the Fibonacci pnumbers – and for the case p=1 to Eq. (4.247) the characteristic equation for the Fibonacci mnumbers. Equation (4.303) has p+1 roots x1, x2, …, xp, xp+1. If we substitute the root xk (k=1, 2, 3, …, p+1) for x into Eq. (4.303), we obtain the following identity for the root xk: (4.304) x kp +1 = mx kp + 1. If we multiply and divide all terms of the identity (4.304) by xk repeated ly, we come to the following general identity (4.305) x kn = mx kn −1 + x kn − p −1 = x k × x kn −1 , where n=0,±1,±2,±3,… and xk (k=1,2,3,…, p+1) is a root of Eq. (4.303). 4.14.4. The Golden (p,m)Proportions According to Descartes’ “rule of signs,” Eq. (4.303) has the only positive root. Suppose without loss of generality that the root x1 is a positive root of Eq. (4.303). Let us denote the root x1 by Φp,m and name it a Golden (p, m) Proportion. Substituting Φp,m for xk (k=1) into the identities (4.304) and (4.305), we obtain two important identities for Φp,m: Φpp,+m1 = mΦpp,m + 1 (4.306) Φpn,m = Φpn,−m1 + mΦpn,−mp −1 = Φp,m × Φpn,−m1 , (4.307) where n=0,±1,±2,±3,…. If we divide all terms of the identity (4.306) by Φpp,m , we obtain the fol lowing remarkable property of the golden (p,m)proportions Φp,m = m + 1 / Φpp,m or (4.308) Φp,m − m = 1 / Φpp,m . (4.309) Alexey Stakhov THE MATHEMATICS OF HARMONY 244 Note that the golden (p,m)proportions Φp,m is a wide generalization of the golden pproportions (m=1), the generalized golden mproportions (p=1) and the classical golden mean (p=1, m=1). Also the identities (4.308) and (4.309) are a wide generalization of the corresponding identities for the golden ppro portions, the golden mproportions, and the classical golden mean. 4.14.5. Properties of the Roots of the Characteristic Equation Theorem 4.4. For a given integer p>0 and a positive real number m, the following correlations for the roots of the golden algebraic equation xp+1mxp1=0 are valid: x1 + x 2 + x 3 + ... + x p + x p+1 = m ( x1 x2 + ... + x1 x p+1 ) + ( x2 x3 + ... + x2 x p+1 )+ ... + ( x p−1 x p + x p−1 x p+1 ) + x p x p+1 = 0 ( x x x + ... + x x x ) + ( x x x + ... + x x x ) + ... + x 1 2 3 1 p p +1 2 3 4 2 p p+1 p−1 (4. 310) x p x p+1 = 0 ... x1 x2 " x p−2 x p−1 x p + x1 x 3 x 4 " x p−1 x p x p+1 + ... + x2 x 3 x 4 "x p−1 x p x p+1 = 0 (4.311) x1 x2 x 3 " x p−1 x p x p+1 = (−1) . p (4.312) Theorem 4.4 is proved by analogy with Theorem 4.1. We use the “Basic Theorem of Algebra” that is given by (4.74) to represent the characteristic equation (4.303) in the form: x p+1 − mx p −1 = ( x − x1 )( x − x2 )"( x − x p+1 ) = x p+1 −( x1 + x 2 + ... + x p+1 ) x p +( x1 x2 + ... + x1 x p+1 + x2 x3 + ... + x2 x p+1 + ... + x p−1 x p+1 + x p x p+1 ) x p−1 −( x1 x2 x3 + ... + x1 x p x p+1 + x2 x3 x4 + ... + x2 x p x p+1 + ... + x p−1 x p x p+1 ) x p−2 +( x1 x2 x 3 x 4 + x1 x 2 x 3 x 5 + ... + x p−2 x p−1 x p x p+1 ) x p−3 (4.313) ... x p−2 x p−1 x p + ... + x2 x 3 x 4 " x p−1 x p x p+1 ) x − x1 x2 x 3 " x p−1 x p x p+1 = 0. +( x1 x2 x 3 "x We can give some explanations regarding the identities (4.310), (4.311), and (4.312) that connect the roots of Eq. (4.303). It is evident from (4.310) that the sum of the roots of Eq. (4.303) is identically equal to m. The expres sion (4.311) gives the values for every possible sum of the roots of Eq. (4.303) taken by two, three, ..., or р roots from the (р+1) roots of Eq. (4.303). Accord ing to (4.311), each of these sums is equal to 0 ! At last, the expression (4.312) gives the value of the product of all roots of Eq. (4.303). According to (4.312) this product is equal to 1 (for the even р) or 1 (for the odd р). Theorem 4.5. For a given integer p=0,1,2,3,… and for the condition, when k takes its values from the set {1,2,3,…, p}, the following identity is valid for the roots of the algebraic equation xp+1mxp1=0: Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 245 ( x + x + x + ... + x + x ) = x + x + x + ... + x + x k k 1 k 2 k 3 k p k k p +1 = m . (4.314) Theorem 4.5 is proven by analogy with Theorem 4.2. For the proof we may consider the following expression: 1 2 3 p p +1 ( x + x + x + ... + x + x ) , k 1 2 3 p (4.315) p +1 where k takes its values from the set {1,2,3,…, p}. Taking into consideration the identity (4.310), we can write: ( x + x + x + ... + x + x ) = m . k k (4.316) On the other hand, if we factorize the expression (4.315) and take into consideration the identities (4.311), we obtain: k ( x1 + x2 + x3 + ... + x p + x p +1 ) = x1k + x2k + x3k + ... + x pk+1 , (4.317) 1 2 3 p p +1 whence the identity (4.314) appears. 4.14.6. Generalized Binet Formulas for the Fibonacci (p,m)Numbers For a given p > 0, the generalized Binet formula for the Fibonacci (p,m) numbers is n n n Fp,m ( n ) = k1 ( x1 ) + k2 ( x2 ) + ... + kp +1 ( x p +1 ) , (4.318) where x1, x2, …, xp+1 are the roots of the characteristic equation (4.303) and k1, k2, …, kp+1 are constant coefficients that depend on the initial terms (4.296) of the Fibonacci (p,m)sequence. In order to calculate the values of the coefficients k1, k2, …, kp+1 in (4.318), we consider solutions of the following system of equations Fp,m (0) = k1 + k2 + " + kp +1 = 0 Fp,m (1) = k1 x1 + k2 x2 + " + kp +1 x p +1 = 1 2 2 2 . Fp,m (2) = k1 x1 + k2 x2 + " + kp +1 x p +1 = m # Fp, m ( p) = k1 x1p + k2 x2p + " + kp +1 x pp+1 = m p −1 (4.319) Solving the system of the equations (4.319), we can obtain the numerical values of the coefficients k1, k2, …, kp+1 for different values of p. For the case p=1, the Fibonacci (p, m)numbers are reduced to the Fi bonacci mnumbers, and we obtain the formula similar to the Gazale formula (4.281), that is, F1,m (n) = Φ1n,m − (−1)n Φ1−,mn 4 + m2 , (4.320) Alexey Stakhov THE MATHEMATICS OF HARMONY 246 where Φ1,m is the golden (1, m)proportion. For the case p=2, the recursive relation (4.295), the seeds (4.296), and the characteristic equation (4.303) take the following forms, respectively, F2,m(n)=mF2,m(n1)+F2,m(n3) (4.321) F2,m(0)=0, F2,m(1)=1, F2,m(2)=m (4.322) (4.323) x3mx21=0. Equation (4.323) has one real root and two complex roots. The roots of Eq. (4.323) are: h2 + 2mh + 4m2 , 6h h m2 h2 − 2mh + 4m2 +i 3 − x2 = − , 6h 12 3h h m2 h2 − 2mh + 4m2 −i 3 − x3 = − , 6h 12 3h x1 = (4.324) (4.325) (4.326) where h = 3 108 + 8m 3 + 12 81 + 12m 3 . The Binet formula for the Fibonacci (2,m)numbers is F2,m ( n ) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x3 ) . n n n (4.327) The values of k1, k2, k3 are solutions of the system F2,m ( 0 ) = k1 + k2 + k3 = 0, F2,m (1) = k1 x1 + k2 x2 + k3 x 3 = 1, F2,m ( 2) = k1 ( x1 ) + k2 ( x2 ) + k3 ( x 3 ) = m. 2 2 2 (4.328) Solving the system, we obtain k1 = 2h(h + 2m) h[−(h + 2m) − i 3 (h − 2m)] 3 3 , k2 = h + 8m h 3 + 8m 3 k3 = h[−(h + 2m) + i 3 (h − 2m)] . h 3 + 8m 3 Taking p=2,3,4,5,… we can derive the generalized Binet formulas for the Fibonacci (p,m)numbers. For example, the Binet formula for the Fibonacci (2, m)numbers is the following: Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 247 2h(h + 2m) h2 + 2mh + 4m2 F2,m (n) = 3 h + 8m 3 6h n h[−(h + 2m) − i 3 (h − 2m)] h2 + 4m2 − 4hm h m2 + i − + 3 12 − 3h h 3 + 8m 3 12h n h[−(h + 2m) + i 3 (h − 2m)] h + 4m − 4hm h m −i 3 − − 3 3 12h h + 8m 12 3h n + 2 2 2 (4.329) 4.14.7. Generalized Binet Formulas for the Lucas (p,m)Numbers By analogy to the formula (4.161), we introduce the following formula that defines the Lucas ( p, m ) numbers: Lp,m ( n ) = ( x1 ) + ( x2 ) + ... + ( x p +1 ) , n n n (4.330) where x1, x2, …, xp+1 are the roots of the characteristic equation (4.303). Calculate the value of the initial (p+1) terms of the number sequence de fined by (4.330). For the case n=0, we have: Lp,m ( 0 ) = ( x1 ) + ( x2 ) + ... + ( x p +1 ) = p + 1. 0 0 0 (4.331) Let us take k from the set {1, 2, 3,..., p} and represent the formula (4.330) in the form: Lp,m ( k ) = ( x1 ) + ( x2 ) + ... + ( x p +1 ) . Using the identity (4.314), we can write: k k k (4.332) Lp,m(k)=mk, (4.333) where k takes its values from the set {1,2,3,…, p}. The expressions (4.331) and (4.333) can be considered as the seeds of the Lucas (p, m)numbers that are given by the following recursive formula: (4.334) Lp,m(n)=mLp,m(n1)+Lp,m(np1). Let us prove that the formula (4.330) gives the same numerical sequence as well as the recursive relation (4.334) at the seeds (4.331) and (4.333). Suppose that the formula (4.330) defines the same number Lp,m(n) that is calculated according to the recursive relation (4.334). Consider the formula (4.330) for the case of n+1, that is, Lp,m ( n + 1) = ( x1 ) n +1 + ( x2 ) n +1 + ... + ( x p +1 ) n +1 . (4.335) Alexey Stakhov THE MATHEMATICS OF HARMONY 248 Using the identity (4.305), we can write the expression (4.335) as follows n n n Lp,m ( n + 1) = m ( x1 ) + ( x2 ) + ... + x p +1 (4.336) − − 1 n p n − p −1 n − p −1 + ( x1 ) + ( x2 ) + ... + x p+1 , ( ( ) ) whence it appears Lp,m(n+1)=mLp,m(n)+Lp,m(np) (4.337) that corresponds to the recursive relation (4.334). This means that our reasoning results in the discovery of a new class of numerical sequences – the mextension of the Lucas pnumbers or simply the Lucas (p, m)numbers. Note that for the case m=1 the recursive relation (4.334) and the seeds (4.331) and (4.333) are reduced to the following: Lp,1(n)=Lp,1(n1)+Lp,1(np1) (4.338) (4.339) Lp,1(0)=p+1 and Lp,1(k)=1 (k=1,2,3,…,p). It is clear that the recursive relation (4.338) at the seeds (4.339) gives the above Lucas pnumbers. For the case m=2, the recursive relation (4.334) and the seeds (4.331) and (4.333) are reduced to the following: Lp,2(n)=2Lp,2(n1)+Lp,2(np1) (4.340) Lp,2(0)=p+1; Lp,2(k)=2k (k=1,2,3,…,p). (4.341) For the case p=1, the Lucas (p,m)numbers are reduced to the Lucas m numbers and we obtain the Binet formula similar to (4.292), that is, (4.342) L1,m ( n ) = Φ1n,m + ( −1) Φ1−,mn , where Φ1,m is the golden (1, m)proportion. For the case p=2, the recursive relation (4.334), the seeds (4.331) and (4.333), the characteristic equation (4.303), and Binet formula (4.330) take the following forms, respectively: n L2,m(n)=mL2,m(n1)+L2,m(n3) (4.343) L2,m(0)=3, L2,m(1)=2, L2,m(2)=4 (4.344) x3mx21=0 (4.345) (4.346) L2. m ( n ) = ( x1 ) + ( x2 ) + ( x3 ) . If we substitute the expressions (4.324)(4.326) for the roots x1, x2, x3 into (4.346), we obtain the following analytical expression of the Binet for mula for the Lucas (2,m)numbers that are called the PellLucas pnumbers: n n n Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 249 n h 2m2 m h m2 m i 3 h 2m2 L2,m (n) = + + + − − + + − 3 12 3h 3 2 6 3h 6 3h n n (4.347) h m 2 m i 3 h 2m 2 + − − + − − . 2 6 3h 12 3h 3 Table 4.8 gives the 2extension of the Lucas pnumbers (PellLucas p numbers) for the cases p=1,2,3,4. Table 4.8. PellLucas pnumbers (m=2, p=1,2,3,4) n 5 4 3 2 1 0 1 2 3 4 5 L1,2(n) 82 34 14 6 2 2 2 6 14 34 82 L2,2(n) 10 8 3 4 0 3 2 4 11 24 52 L3,2(n) 0 4 6 0 0 4 2 4 8 20 42 L4,2(n) 5 8 0 0 0 5 2 4 8 16 37 It is clear that our reasoning resulted in a wide generalization of the Fibonacci and Lucas pnumbers and the Fibonacci and Lucas mnumbers. The Fibonacci and Lucas (p,m)numbers are of theoretical interest for discrete mathematics and open new perspectives for the development of theoretical physics because new recursive numerical sequences and new mathematical constants that follow from this approach may be discovered in some physical processes. 4.15. Structural Harmony of Systems 4.15.1. Soroko’s Law of the Structural Harmony of Systems Belorussian philosopher Eduardo Soroko was one of the first researchers who used the Fibonacci pnumbers and golden pproportions for simulation of the processes in selforganizing systems [25, 56]. Soroko’s main idea is to con sider real systems from the dialectical point of view. As is well known, any nat ural object can be represented as the dialectical unity of the two opposite sides A and B. This dialectical connection may be expressed in the following form: A+B=U (universum). (4.348) Alexey Stakhov THE MATHEMATICS OF HARMONY 250 The equality (4.348) is the most general expression of the socalled Conservation Law. Here A and B are distinctions inside of the Unity, logically disjoint classes of the Whole. There is one requirement that A and B need to be measured by the same measure and members of the relation that underlies the unity. Probability and improbability of events, mass and energy, the nu cleus of an atom and its envelope, substance and field, anode and cathode, animals and plants, spiritual and material beginnings in a value system, and profit and cost are various examples of (4.348). The identity (4.348) may be reduced to the following normalized form: (4.349) A + B = 1, where A and B are the relative “weights” of the parts A and B that make up some Unity. The Law of Information Conservation is a partial case of (4.348): I+H=logN, (4.350) where I is an information quantity and H is the entropy of the system having N different states. The normalized form of the law (4.350) is the following: (4.351) R + H = 1, where R=1/logN is a relative redundancy, H = H log N is a relative entropy. Let us consider the process of system selforganization. This one is re duced to the passage of the system into some “harmonious” state called the state of thermodynamic equilibrium. There is some correlation or proportion between the sides A and B of the dialectical contradiction (4.348) for the state of thermodynamic equilibrium. This correlation has a strictly regular charac ter and is the cause of system stability. Soroko uses the Principle of Multiple Relations to find a connecting law between A and B in the state of thermody namic equilibrium. This principle is well known in chemistry as Dalton’s Law and in crystallography as the Law of Rational Parameters. Soroko proposes the hypothesis that the Principle of Multiple Relations is a general principle of the Universe. That is why, in accordance with this prin ciple there is the following connection between the components R and H in the identity (4.351): log R = ( s + 1) H or (4.352) (4.353) log H = ( s + 1) R. The equalities of (4.352) and (4.353) may be represented in exponential form: Chapter 4 ( ) R= H Generalizations of Fibonacci Numbers and the Golden Mean 251 s +1 (4.354) (4.355) H = R s +1 , where the number s is called the Range of Multiplicity and takes the following values: 0,1,2,3,… . Substituting the expressions (4.354) and (4.355) into the equality (4.351), we obtain the following algebraic equations, respectively: (H ) s +1 + H −1 = 0 (4.356) (4.357) R s +1 + R − 1 = 0. If we denote the variables H and R in the equations (4.356) and (4.357) by y, we obtain the following algebraic equation: ys+1+y1=0. (4.358) Let us introduce the new variable x=1/y and apply it to the equation (4.358). Inserting the new variable into (4.358), we obtain the following alge braic equation: (4.359) xs+1+xs1=0. We can see that the equation (4.359) coincides with the algebraic equa tion of the golden pproportion that is given by (4.42). The positive root of the equation (4.358) is the reciprocal of the golden pproportion: (4.360) βs=1/τs, where τs is a positive root of the equation (4.359). In accordance with Soroko’s concept, the roots of the equation (4.358) that are equivalent to equation (4.359), express the Law of the Structural Har mony of Systems: “The generalized golden proportions are invariants that allow natu ral systems in the process of their selforganization to find harmonious structure, a stationary regime for their existence, and structural and func tional stability.” What is peculiar about Soroko’s Law? Since Pythagoras, scientists combined the concept of Harmony solely with the classical golden mean τ = 1 + 5 2. Soroko’s Law asserts that the harmonious state that corre sponds to the classical golden mean is not unique to a system. Soroko’s Law asserts that there are an infinite number of harmonious states of the system that correspond to the numbers τ s or to the inverse numbers β s=1/ τ s (s=1,2,3,…) that are positive roots of the general algebraic equations (4.358) and (4.359). ( ) Alexey Stakhov THE MATHEMATICS OF HARMONY 252 Table 4.9 gives the values of Soroko’s “structural invariants” for the initial terms of s. Table 4.9. Soroko’s numerical invariants s βs 1 2 3 4 5 6 7 0.618 0.682 0.724 0.755 0.778 0.796 0.812 4.15.2. Application of Soroko’s Law to Thermodynamic and Information Systems The thermodynamic or informational state of a system is expressed by en tropy, which is the principal concept of thermodynamics and information theo ry. The expression for the entropy of an information source with the alphabet A={a1, a2, …, aN} has the following form: N H =− ∑ p log p , k (4.361) k k =1 where p1, p2, …, pN are the probabilities of the letters a1, a2, …, aN, N is a number of the letters. It is well known that entropy (4.361) reaches its maximum value (4.362) Hmax=logN for the case, when the probabilities of the letters are equal among themselves, that is, p1=p2= …= pN=1/N. Using the concept of relative entropy H = H log N , we can write the following evident equality: (4.363) N H log N = H = − ∑ p log p . k k (4.364) k =1 In accordance with the Law of the Structural Harmony of Systems, any system reaches its harmonious state when its relative entropy (4.363) satis fies the equality (4.353). The following expression for the entropy of the harmonious system follows from this consideration: N H =− ∑ p log p = β log N . k k s (4.365) k =1 It is clear that for any given s, the problem of obtaining the optimal set of values pi (i=1, 2, 3, …, N) corresponding to the optimal (“harmonious”) state Chapter 4 Generalizations of Fibonacci Numbers and the Golden Mean 253 of the system, has many solutions. However, the expression (4.365) plays a role in the function’s “aim” towards the solution of various scientific and technical problems. In his book, Structural Harmony of Systems [25], Soroko gave a number of interesting examples from different fields of science to confirm the Law of the Structural Harmony of Systems. For example, let us consider such a natural object as “dry air” that is the basis for life on Earth. We can ask the question: does the “dry air” have an optimal or harmonious structure? Soroko’s theory gives a positive answer to this question. In fact, the chemical compound of the “dry air” is the following: nitrogen – 78.084%; oxygen – 20.948%; argon – 0.934%; carbon dioxide – 0.031%; neon – 0.002%; and helium – 0.001%. If we calculate the entropy of the “dry air” according to the formula of (4.361) and then its relative entropy according to (4.363) taking into consideration that logN=log 6, then the value of the relative entropy of the “dry air” will be equal to 0.683. With a high degree of precision this value corresponds to the invari ant β2=0.682. This means that in the process of selforganization the “dry air” reaches its optimal, harmonious structure. This example is very typical and demonstrates that “Soroko’s Law” can be used today for checking the bio sphere state, in particular, the states of air and water. It is clear thus the practical use of the Law of the Structural Harmony of Systems provides a clear advantage to the solution of many technological, eco nomical, ecological and other problems. In particular, it can help improve the technology of structurallycomplicated products, including the monitoring of the biosphere and so forth. 4.16. Conclusion In recent years the generalization of Fibonacci and Lucas numbers and the golden section has been an important trend in the development of Fi bonacci number theory. One may cite various ways in which this generaliza tion takes place. One way, for example, is a generalization based on Pascal’s triangle. In the latter half of the 20th century several eminent mathemati cians (including Martin Gardner [12], George Polya [17], Alfred Renyi [23] and others) each independently discovered a connection of Fibonacci num bers with Pascal’s triangle and binomial coefficients. This finding confirms a fundamental connection of Fibonacci and golden mean based Harmony Math ematics with combinatorial analysis suggesting a future line of development. Alexey Stakhov THE MATHEMATICS OF HARMONY 254 In the early 1970s, in his doctoral dissertation [19], Alexey Stakhov intro duced the socalled Fibonacci pnumbers (p=0,1,2,3,…) that follow from the diagonal sums in Pascal’s triangle. By studying the Fibonacci pnumbers, Sta khov generalized the golden section problem and introduced the golden p proportions that are positive roots of the characteristic equation xp+1=xp+1. Later by continuing this research, Stakhov formulated the Generalized Prin ciple of the Golden Section [107] and generalized the Euclidean problem of the “division in extreme and mean ratio” (the golden section) [20]. Later Stakhov and Rozin generalized the Binet formulas and introduced the generalized Lucas pnumbers [111]. Fibonacci numbers were further generalization through a consideration of the generalized recursive relation G(n+1)=pG(n)+qG(n1) (p and q are integers or real numbers) that coincides with the classical recursive Fibonac ci relation for the case p=q=1. This became the source of many original dis coveries made by Vera W. de Spinadel [42], Midhat Gazale [45] and Jay Kap praff [47]. By using this approach, Vera W. de Spinadel introduced the Me tallic Means [42]. By using the recursive relation Fm(n+1)=mFm(n)+Fm(n1) and the characteristic equation x2mx1=0, where m is a positive real number, Midhat Gazale further generalized the Binet formula for Fibonacci mnum bers Fm(n). Gazale formulas allow one to represent all Fibonacci and Lucas m numbers by the golden mproportion Fm that is a generalization of the classi cal golden mean. Recently, E. Gokcen, Naim Tuglu and Alexey Stakhov [154] introduced the mextension of the Fibonacci and Lucas pnumbers, generating the new class of Fibonacci and Lucas (p,m)numbers. This new class of characteristic equations xp+1mxp1=0, for generalized Fibonacci and Lucas (p,m)numbers and the golden mean, infinitely extend algebraic equations and mathematical constants, which can be used in contemporary mathematics, theoretical phys ics and computer science. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 255 Chapter 5 Hyperbolic Fibonacci and Lucas Functions 5.1. The Simplest Elementary Functions 5.1.1. Trigonometric Functions In mathematics, an Elementary Function is built up from a finite number of exponentials, logarithms, constants, variables, and roots of equations by using the four elementary operations (addition, subtraction, multiplication and division). Logarithms, Exponential Functions (including Hyperbolic Func tions), Power Functions, and Trigonometric Functions are the best known amongst them. We begin to study the elementary functions starting with trigonometric functions. Trigonometry and trigonometric functions Sine, Cosine, Tangent, Co tangent, Secant, and Cosecant are well known to many of us from secondary school. Trigonometry was developed in antiquity initially as a branch of astron omy, as a computing tool for practical purposes. Application of trigonometry to astronomy explains why spherical trigonometry arose first, before planar trigo nometry. Ancient Greek astronomers successfully solved some problems of trig onometry. However, these scientists (Ptolemy, Menelaus, etc.) had not studied trigonometric functions (sine, cosine, etc.) but line segments and spans. The span that connects together a double arch 2α played a role of a sine of the angle α. Ptolemy had found a formula for the definition of the span as the sum and differ ence of two arches, the span of the half arch, and so on. Sine and cosine were first introduced by Indian scientists. In India, the doctrine of trigonometric values was named Goniometry as it started to de velop. The further development of the doctrine of trigonometric values con tinued in the 9th15th centuries in the countries of the Middle and the Near East. The Arabian mathematicians introduced all basic trigonometric func tions derivative from sine and cosine: tangent, cotangent, secant and cose cant. They had proved the basic relations amongst trigonometric functions. Alexey Stakhov THE MATHEMATICS OF HARMONY 256 For the first time Arabian mathematicians began to develop trigonometry as a part of mathematics independent from astronomy. In the 13th15th centuries, the further development of trigonometry in Eu rope continued following the translation of mathematical and astronomical works of Arabian and Greek science into Latin. The German scientist Regiomontanus (14361476) was the most famous European mathematician of this epoch in the field of trigonometry. His trigonometric work Five Books about Triangles of All Kinds was of great importance for the further development of trigonometry dur ing the 16th17th centuries. On the threshold of the 17th century, an analytical direction in trigonometry started to develop. Before the 17th century, the direction regarding triangles and the calculation of the elements of geometric figures was the main objective of trigonom etry and the doctrine of trigonometric functions had developed along geometrical lines. Whereas in the 17th19th centuries, trigonometry gradually became one of the branches of mathematical analysis. Trigonometry finds wide application in mechan ics, physics and engineering, especially in the study of oscillatory movements and other periodic processes. The development of the doctrine of oscillatory movements, of sound, light and electromagnetic waves became central to the basic contents of trigonometry. It is well known from physics that the equation of harmonious fluctu ation (for example, of a variable electric current) appears as: y = A sin ( ωt + α ) . Sinusoids are graphs of harmonious fluctuations; therefore, in physics and engineering the harmonious fluctuations are often called Sinusoidal Fluctuations. Newton and Euler developed an analytical approach to trigonometry, and in the first half of the 19th century the French scientist Fourier proved that any periodic movement can be presented (with any degree of accuracy) in the form of the sum of simple harmonious fluctuations. Presently trigonometry is no longer considered an independent part of mathematics. Its major sec tion – the Doctrine about Trigonometric Function – is a part of the theory of functions; and its other section –the Decision of Triangles – is considered to be part of geometry (planar and spherical). 5.1.2. The Power Function The Power Function plays an important role in mathematics. The Power con cept originally meant the product of a finite number of equal coefficients (a pow er with a natural parameter), that is, an=a×a×...×a (n times). (5.1) Chapter 5 Hyperbolic Fibonacci and Lucas Functions 257 This definition possesses the following mathematical properties: ( ab ) = a n × bn ; ( a / b ) = a n / bn ; n n ( am ) = amn ; am / an = am−n . n (5.2) Over the centuries, this concept was repeatedly generalized and enriched by its contents. The concept of the 2nd and 3rd powers a2 and a3 appeared in connection with the definition of the area of a square and the volume of a cube. Already the Babylonians had made use of tables of the squares and cubes of numbers. The titles Square and Cube for the 2nd and 3rd powers are of ancient Greek origin. The practice of solving more and more complex algebraic problems, led to the necessity to generalize the power concept and its extension by means of the introduction of the powers of zero, and negative and fractional numbers. The powers of zero and also negative and fractional parameters were defined so that the actions of the power with a natural basis given by (5.2) were applied to them. The principle here ob served in the generalization of mathematical concepts is referred to as “a permanence principle.” The power concept (5.1) depends on two parameters, the numbers of a and n. Depending on what parameter is chosen as a variable, it is possible to give two gener alizations of the power concept (5.1). The first generalization is the power function: y=xσ, (5.3) where x is a variable and σ is a given real number. Note that for the cases σ=0 and σ=1 according to (5.3) we have, respectively: y=1 and y=x. (5.4) The graph of the function y=1 is a straight line parallel to the axis OX, and the graph of the function y=x is a bisector of the 1st and 3d coordinate angles. For the case σ=2 the graph of the function (5.3) is a Parabola y=x2. For the case σ=3 the graph of the function (5.3) is a Cubic Parabola y=x3. This curve had been used by French mathematician Gaspard Monge (17461818), the father of descriptive geometry, for finding the real roots of the cubic equations. A derivative of the power function (5.3) is equal to the product of the power parameter α multiplied by the power function with the parameter (α1), that is, d α x = αx α −1 . dy The power function (5.3) is a basic mathematical tool that is used widely in many other fields as well, including economics, biology, chemistry, physics, and computer science, with such applications as compound interest, population growth, chemical reactions, wave behavior, and public key cryptography. Alexey Stakhov THE MATHEMATICS OF HARMONY 258 5.1.3. An Exponential Function However, there is another generalization of the power concept (5.1). Let us examine the socalled exponential function given by the following expression: (5.5) y=ax, where x is a variable and a>0 is any real number. The exponential function (5.5) has many applications in the study of natu ral and social phenomena. It is known, for example, that the decay of radioactive substance is described by the function (5.5). If we denote by t0 a period of the halfdecay, that is, a time interval necessary to get half of the initial mass m0, then the mass of the substance via t years can be expressed as follows: d α x = αx α −1 . dy The exponential functions (5.5) posses the following Exponential Laws: a0 = 1; a1 = a; a x + y = a x a y ; a xy = ( a x ) ; y 1 / a x = a − x ; a x b x = ( ab ) . x These correlations are valid for all positive real numbers a and b and all values of the variables x and y. The expressions that involve fractions and roots can often be simplified by using exponential functions because: 1 a = a −1 ; n ( ) ab = n a b b = an , where n is a natural number. In mathematics, the following partial case of the exponential function is widely used: y = ex , (5.6) where e=2.71828183 is the base of natural logarithms. The graph of the exponen tial function (5.6) is represented in Fig. 5.1. y 5 The exponential function is climbing 4 slowly with the negative values of x, it is climbing quickly with the positive values 3 of x, and is equal to 1 when x is equal to 0. 2 As a function (5.6) is the function of real variable x, the graph of (5.6) is positive for 1 1 2 all values of x and is increasing (by viewing 5 4 3 2 1 x lefttoright). This graph never touches the 1 xaxis, although it approaches the xaxis; thus, the xaxis is a horizontal asymptote 2 to the graph. Figure 5.1. Exponential function Chapter 5 Hyperbolic Fibonacci and Lucas Functions 259 A derivative of the exponential function (5.6) coincides with the function (5.6), that is, d x e = ex . dx Note that only the function (5.6) has this unique property. For the exponential functions with other bases we have: d x a = (ln a)a x . dx 5.1.4. Logarithms The theory of logarithms is based on the following simple reasoning. Let us examine the exponential function (5.5). First we represent the variable x in (5.5) in the form of a special function of y as follows: ( ) x = log a y = log a a x . We name this function a Logarithm with base a. Depending upon the choice of the base a there are different kinds of logarithms. For the case a=2 we have Binary Logarithms log2x=lgx, for the case a=10 we have Decimal Logarithms log10x and finally, for the case a=e we have Natural Logarithms logex=lnx. Logarithms are used widely in many areas of science and engineering where the magnitudes vary over a large range. Logarithmic scales are used, for example, in the decibel scale for the loudness of sound, the Richter scale of earthquake magnitudes, and the astronomical scale of stellar brightness. A derivative and an indefinite integral of logaz are given respectively by d 1 z(ln z − 1) log a z = ; log a dz = + c. dz z ln a ln a ∫ 5.2. Hyperbolic Functions 5.2.1. Definition of the Hyperbolic Functions The function e x − e− x sh( x ) = 2 is called the Hyperbolic Sine and the function (5.7) Alexey Stakhov THE MATHEMATICS OF HARMONY 260 e x + e− x 2 is called the Hyperbolic Cosine. ch( x ) = y y = ch(x) (5.8) y = sh(x) There is a similarity between trigono metric and hyperbolic functions. Like the 4 trigonometric sine and cosine that are the 3 coordinates of the points on a circle, the hy 2 perbolic sine and cosine are the coordinates 1 of the points on a hyperbola. The hyper 4 3 2 1 0 1 3 5 2 4 x bolic functions are defined on all numeri 1 cal axes. The hyperbolic sine is an odd func 2 tion that is increasing on all numerical axes. 3 The hyperbolic cosine is an even function 4 that is decreasing on the interval (∞; 0) and increasing on the interval (0; +∞). The Figure 5.2. Graphs of the point (0; 1) is the minimum of this func hyperbolic sine and cosine tion (see Fig. 5.2). By analogy with the trigonometric functions we can define hyperbolic tan gent and cotangent: sh( x ) ch( x ) th( x ) = , cth( x ) = . ch( x ) sh( x ) y 5 Analytical definitions (5.7) and (5.8) of the hyperbolic functions can be used for obtaining some important identities of the hyperbolic trigonometry. It is well known that there are Trigonometric Identities, for example, the Pythagore an Theorem for the Trigonometric Functions: cos2a+sin2a=1. We can prove similar identities for the hyperbolic functions: 2 e x + e− x e x − e− x ch x − sh x = − 2 2 −2 x −2 x 2x 2x e +2+e e −2+e =1 = − 4 4 2 2 2 2 (5.10) 2 e x + e− x e x − e− x ch x + sh x = + 2 2 e2 x + 2 + e −2 x e2 x − 2 + e −2 x e2 x + e −2 x = = sh2 x . = + 4 4 2 2 (5.9) 2 (5.11) Chapter 5 Hyperbolic Fibonacci and Lucas Functions 261 We can prove the following properties for derivatives and integrals: ( shx ) ′ = chx ; ( chx ) ′ = shx ; ( thx ) ′ = 1 / ch2 x ∫ sh( x )dx = ch( x ) + C ∫ ch( x )dx = sh( x ) + C ∫ th( x )dx = ln[ch( x )] + C 1 ∫ ch ( x )dx = th( x ) + C 2 1 ∫ sh ( x ) dx = −cth( x ) + C 2 5.2.2. History and Applications of the Hyperbolic Functions 5.2.2.1. Lambert and Riccati Although Johann Heinrich Lambert (17281777), a French mathemati cian, is often credited with introducing hyperbolic functions, it was actually Vincenzo Riccati (17071775), an Italian mathematician, who did this in the middle of the 18th century. He studied these functions and used them to obtain solutions for cubic equations. Riccati found the standard Addition for mulas, similar to trigonometric identities, for hyperbolic functions as well as their derivatives. He revealed the relationship between the hyperbolic func tions and the exponential function. For the first time Riccati used the sym bols sh and ch for the hyperbolic sine and cosine. 5.2.2.2. NonEuclidean Geometry Among the mathematical works of ancient science, The Elements by Euclid is of special importance. This famous work contains the fundamentals of ancient mathematics: elementary geometry, number theory, algebra, the general theory of relations and calculation of areas and volumes. Euclid summarized Greek math ematics and created a stable foundation for further mathematical progress. The Elements by Euclid is constructed as a deductive mathematical system: at first the definitions, the postulates and the axioms are given, then the theorems are formu lated and their proofs are given. Among the Euclidean axioms, the 5th Euclidean axiom, the Axiom about Parallels is the most famous. During almost 2 millennia many mathematicians tried to deduce this axiom as a theorem from other Euclidean axioms. The problem of the 5th Euclidean Alexey Stakhov THE MATHEMATICS OF HARMONY 262 axiom was brilliantly solved for the first time by a professor of Kazan University, the Great Russian mathematician Nikolay Lobachevsky (17921856) who devel oped a nonEuclidean geometry in 1826. Lobachevsky’s geometry is also named Hyperbolic Geometry because it is based on the hyperbolic functions. Indepen dently Lobachevsky, the young Hungarian mathematician Janosh Bolyai had also developed a similar nonEuclidean geometry. The first published work on nonEuclidean geometry, Lobachevsky’s article About the Geometry Beginnings, was published in 1829 in The Kazan Bulletin. Three years later Janosh Bolyai’s work on nonEuclidean geometry, called the Appendix, was published in Latin. After Gauss’ death it was clear that he also had developed a geometry similar to those of Lobachevsky and Bolyai. Lobachevsky had formulated a new axiom about parallel lines that, in contrast to the Euclidean 5th axiom, was formulated as follows: “Through any point outside of a given straight line, it is possible to draw at least two different ‘parallel’ straight lines that are mutually disjoint with the given straight line.” Having replaced the 5th Euclidean axiom with this new axiom, Lobachevsky developed a nonEuclidean geometry that is as logically correct as Euclidean geometry. We will not stop to detail all features of Lobachevsky’s geometry, as its study goes far outside the limits of “Elementary mathematics,” studied in high school. It is important to emphasize that, by studying trigonometric relations of his geometry, Lobachevsky used the above hyperbolic functions, that is, Lobachevsky’s geometry is an important confirmation of the funda mental character of hyperbolic functions in the development of new geomet ric models of the Universe. 5.2.2.3. Minkowski’s Fourdimensional World The 20th century became a new stage in the evolution of spatial ideas in all scientific spheres. Physics stimulated the global process of change regarding geometric spatial ideas. Until the creation of nonEuclidean geometry, it was not necessary to prove the mutual relation of Newton’s mechanics and Euclid ean geometry. This fact was considered to be obvious. However, at the end of the 19th century and the beginning of the 20th century many new physical facts and observations, which did not fit the classical spatial representations, were gathered together. Maxwell’s research on electrodynamics, Michelson’s experiments on the measurement of the speed of light and other scientific facts resulted in the problem of the correspondence of Euclidean geometry to the real physical world. Einstein’s theory of relativity resulted in the explanation of Chapter 5 Hyperbolic Fibonacci and Lucas Functions 263 new physical facts, which concern relations between physics and classical geome try. The conclusion about the nonEuclidean character of real spatial geometry became the main outcome of Einstein’s special theory of relativity. The need for attracting new geometric ideas arose within 20th century phys ics. In particular, Einstein’s special relativity theory (1905) was a cause for this research. In this theory for the first time in the history of physics, a mutual rela tion between space and time became the objects of physical research. Relativity theory proves that the metrics of real space and time are not absolute: they de pend on dynamical conditions, in which the spatial and temporal measurements are carried out. Relativity theory for the first time showed that space and time are continuum and that their properties are integrally interconnected. In 1908, that is, in three years after the promulgation of Einstein’s special theory of relativity, the German mathematician Herman Minkowski presented the geometric substantiation of the special relativity theory [37]. Minkowski’s idea is characterized by two essential peculiarities. In the first place, the geomet ric spatialtemporal model that was offered by him was fourdimensional: in this model the spatial and temporal coordinates are connected in the common coor dinate system. The position of the material point in Minkowski’s model was de termined by the point M (x, y, z, t) called the World Point. Secondly, the geo metric connection between the spatial and temporal coordinates in Minkows ki’s system had a nonEuclidean character, that is, the given model reflected cer tain special properties of real spacetime that could not be simulated in the frame work of the “traditional” Euclidean geometry. Geometrically, a connection between the spatial (x) and temporal (t) co ordinates in Minkowski’s model is given by the Hyperbolic Turn, the move ment that is similar to the traditional turn of the Cartesian system in Euclid ean geometry. Here, the coordinates x and t of any point can be transformed according to the formulas: x ′ = xchψ + tshψ, t ′ = xshψ + tchψ, where ψ is the angle of the hyperbolic turn and ch and sh are the hyperbolic sine and cosine, respectively. The original geometric interpretation of the wellknown Lorentz transfor mations follows from the given model: x ′ = ( x + vt ) ( 1− (v c) ; 2 ) t ′ = t + vx c 2 1− (v c) , 2 where v is the speed of the system, с is the speed of light. Alexey Stakhov THE MATHEMATICS OF HARMONY 264 Note that in Minkowski’s geometry Lorentz transformations are rela tions of the hyperbolic trigonometry expressed in physics terms. Minkows ki’s geometry uncovers the hyperbolic nature of all mathematical formulas of relativity theory. 5.2.2.4. Vernadsky’s Ideas Up until the time of relativity theory the mathematical apparatus for biological research practically did not exist. Nevertheless the response of biologists to discoveries in geometric physics was practically instantaneous. Russian scientist Vladimir Vernadsky was one of the first scientists to give serious attention to the geometric problems of biology [37]. Vernadsky’s pecu liar approach was a holistic vision of the “spacetime” problem. Vernadsky ex panded on the concepts of the “spacetime” problem. He took into consideration the biological specificity of its study and discussed the realization in Nature of general geometric (spatialtemporal) laws. Vernadsky gave attention to the prob lem of biological symmetry and considered this to be the key biological problem. He came to the conclusion that the explanation of biological symmetry is con nected to the nonEuclidean character of the spatial geometry of living sub stance. He also assumed that the geometry of biological “spacetime” differs from the geometry of “spacetime” relativity theory, known as the “Fourdimensional Minkowski world” [37]. According to his opinion, the issue of how nonEuclidean geometry is embodied in living nature is the key problem of biological research. Vernadsky’s ideas became a leitmotiv for scientific research in subsequent periods of development in biology resulting in the creation of a new direction, mathematical biology. 5.3. Hyperbolic Fibonacci and Lucas Functions (StakhovTkachenko’s Definition) 5.3.1. A Definition of Hyperbolic Fibonacci and Lucas Functions Fibonacci and Lucas functions were introduced, for the first time, in 1993 by Alexey Stakhov and Ivan Tkachenko [98]. If we compare the hyperbolic func tions (5.7) and (5.8) with Binet formulas (2.66) and (2.67), we can see that Binet formulas are similar to the hyperbolic functions in their mathematical structure. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 265 For a strong definition of the hyperbolic Fibonacci and Lucas functions we can rewrite the formulas (2.68) and (2.67) respectively, as follows: ( F2 k = τ2 k − τ−2 k ( F2 k +1 = τ 2 k +1 +τ 2 k +1 −τ L2 k = τ + τ −2 k L2 k +1 = τ 2k ) 5 −( 2 k +1) (5.12) ) 5 − ( 2 k +1) (5.13) (5.14) , (5.15) where the discrete variable k takes its values from the set {0, ±1, ±2, ±3,...} . Comparing the formulas (5.12)(5.15) to the hyperbolic functions (5.7) and (5.8), we can see that formulas (5.12) and (5.14) correspond in their structure to the hyperbolic sine (5.7) and the formulas (5.13) and (5.15) correspond to the hyperbolic cosine (5.8). This simple analogy underlies the Hyperbolic Fibonacci and Lucas Functions [98]. By substituting the discrete variable k in the formulas (5.12)(5.15) for the continuous variable x that takes its values from the set of real numbers, we obtain the following four new elementary functions: Hyperbolic Fibonacci sine sF ( x ) = ( τ2 x − τ −2 x ) 5 . (5.16) Hyperbolic Fibonacci cosine − 2 x +1 cF ( x ) = τ2 x +1 + τ ( ) 5. ( ) (5.17) Hyperbolic Lucas sine sL( x ) = τ2 k +1 − τ−(2 k +1) . Hyperbolic Lucas cosine (5.18) (5.19) cL( x ) = τ2 k + τ−2 k . Note that for the discrete values x=k the hyperbolic Fibonacci and Lucas functions (5.16)(5.19) coincide with the Fibonacci and Lucas numbers, that is, sF ( k ) = F2 k ; cF ( k ) = F2 k +1 ; sL ( k ) = L2k +1 ; cL ( k ) = L2k . (5.20) This means that the extended Fibonacci and Lucas numbers coincide with the hyperbolic Fibonacci and Lucas functions for the discrete points of the continuous variable x=k (k=0,±1,±2,±3,…), that is, the extended Fibonacci and Lucas numbers are “discrete analogs” of the hyperbolic Fibonacci and Lucas sine and cosine. This property (5.20) is a rather characteristic peculiarity of the above hyperbolic Fibonacci and Lucas functions (5.16)(5.19) in compar ison with the classical hyperbolic functions (5.7), (5.8) that, however, do not have “discrete analogs” similar to (5.20). Alexey Stakhov THE MATHEMATICS OF HARMONY 266 5.3.2. The Hyperbolic Fibonacci and Lucas Tangent and Cotangent Let us introduce the definitions of the hyperbolic Fibonacci and Lucas tangents and cotangents by analogy to the classical hyperbolic tangent and cotangent. Hyperbolic Fibonacci tangent: sFx tFx = . (5.21) cFx Taking into consideration (5.16) and (5.17), the hyperbolic Fibonacci tan gent can be represented as follows: tFx = τ2 x − τ −2 x (τ4 x − 1)τ2 x +1 τ(τ4 x − 1) = 2 x 4 x +2 = . − (2 x +1) +τ + 1) τ4 x +1 + 1 τ τ (τ 2 x +1 (5.21) Hyperbolic Fibonacci cotangent: cFx . sFx Taking into consideration (5.16) and (5.17), the expression for the hyper bolic Fibonacci cotangent can be represented as follows: τ4 x + 2 + 1 ctFx = . (5.22) τ(τ4 x − 1) ctFx = Hyperbolic Lucas tangent: sLx τ4 x + 2 − 1 . tLx = = cLx τ(τ4 x + 1) (5.23) Hyperbolic Lucas cotangent: cLx τ(τ4 x + 1) = . (5.24) sLx τ4 x + 2 − 1 Thus, the main result that follows from our study is the introduction of two new kinds of elementary functions, the Hyperbolic Fibonacci Functions and the Hyperbolic Lucas Functions. By their form, these functions are similar to the classical hyperbolic functions, but differ from them by one important feature. In contrast to the classical hyperbolic functions, the hyperbolic Fibonacci and Lu cas functions have numerical analogs the classical Fibonacci and Lucas num bers. In particular, the Fibonacci numbers with the even indices are discrete analogs for the hyperbolic Fibonacci sine (5.16), the Fibonacci numbers with the odd indices are discrete analogs for the hyperbolic Fibonacci cosine (5.17), the Lucas numbers with the odd indices are discrete analogs for the hyper bolic Lucas sine (5.18), and the Lucas numbers with the even indices are discrete analogs for the hyperbolic Fibonacci cosine (5.19). ctLx = Chapter 5 Hyperbolic Fibonacci and Lucas Functions 267 5.3.3. Some Properties of the Hyperbolic Fibonacci and Lucas Functions 5.3.3.1. Hyperbolic Fibonacci Sine The function y=sF(x) is odd function because ( sF ( − x ) = τ−2 x − τ−( −2 x ) ) ( 5 = − τ2 x − τ−2 x ) 5 = − sF ( x ) . Its value at the point x=0 is equal to ( sF ( 0 ) = τ0 − τ0 ) 5 = 0. 5.3.3.2. Hyperbolic Fibonacci Cosine The function y=sF(x) is given by (5.17). Let us find the intersec tion points of the function graph with the coordinate axes. As cF ( x ) = τ2 x +1 + τ −(2 x +1) 5 > 0 for any x∈{∞,+∞} this means that the graph of the function does not intersect the xaxes at any point. For x=0 we have ( ( cF ( 0 ) = τ1 + τ−1 ) ) 5 = 1, that is, the graph of the function y=sF(x) intersects the yaxis at the point y=1. Now, let us examine the function y=cF(x) on the extremum; for this pur pose we find its first derivative: 2 ln τ 2 x +1 −(2 x +1) 2 y′ = (cFx )′ = τ −τ = ln τ sLx . 5 5 Note that y′=0 for the condition: ( ) τ2 x +1 − τ ( ) = 0; τ4 x + 2 = 1; 4 x + 2 = 0; x = −1 2 . Thus, we have the extremum at the point with abscissa x=1/2. It is easy to prove that the function y=cF(x) is symmetric about the line x=1/2 because − 2 x +1 cF x − (1 2 ) = ( τ2 x + τ −2 x ) 5 = cF − x − (1 2 ) . It is important to note that the function y=cF(x) is not symmetric about the yaxis. 5.3.3.3. Hyperbolic Lucas Sine The hyperbolic Lucas sine y=sL(x) is given by (5.18). Let us calculate the value of the function at the point x=0: ( sL ( 0 ) = τ − τ−1 = 1 + 5 ) 2 + (1 − 5 ) 2 = 1. Alexey Stakhov THE MATHEMATICS OF HARMONY 268 Now, let us calculate the value of x where the function y=sL(x) is equal to 0: τ2 x +1 − τ ( ) = 0; τ4 x + 2 = 1; 4 x + 2 = 0; x = −1 2 . As sL(1/2)=0, this means that the graph of the function y=sL(x) intersects the xaxis at the point x = −1 2. − 2 x +1 5.3.3.4. Hyperbolic Lucas Cosine The function y=cL(x) is given by (5.19). By using the calculus methods [154], we can find the following properties of the function y = cL(x). The function is an even function because cL ( − x ) = τ−2 x + τ ( ) = τ−2 x + τ2 x = cLx . Let us find the intersection points of the function graph with the coordinate axes. As cL(x)=τ2x+τ2x>0 for any x∈{∞,+∞}, this means that the graph of the function does not intersect the xaxes at any point. For the case x=0 we have: cL(0)=τ0+τ0=2; therefore, the graph of the func tion intersects the yaxis at the point y=2. Let us examine the function y=cL(x) on the extremum; with this purpose we find the first derivative of the function: y′ = cL ( x ) ′ = 2 ln τ ( τ2 x − τ −2 x ) . − −2 x The first derivative y′ turns out to be 0 only under the condition that τ2xτ2x=0, this is possible only for the case x=0. This means that at the point x=0 the function has an extremum. The extreme value of the function y=cL(x) at the point x=0 is equal to cL(0)=2. By analogy, we can examine other hyperbolic Fibonacci and Lucas functions, in particular, tangent and cotangent, secant and cosecant. 5.4. Integration and Differentiation of the Hyperbolic Fibonacci and Lucas Functions and their Main Identities 5.4.1. Integration of the Function y=sF(x) Represent the function y=sF(x) given by (5.16) in the form: sF ( x ) = ( τ2 x − τ −2 x ) 5 = ( e2 x ln τ − e −2 x ln τ ) 5 . (5.25) Then, by using the representation (5.25), we obtain the following expres sion for the integral of the function y=sF(x): Chapter 5 Hyperbolic Fibonacci and Lucas Functions 269 ∫ sF ( x ) dx = ∫ ( e − e = (1 5 ) ( ∫ e dx − ∫ e 2 x ln τ 2 x ln τ −2 x ln τ −2 x ln τ ) dx ) 5 dx ( 5 ) (1 2 ln τ ) e + (1 2 ln τ) e = (1 2 5 ln τ ) ( e +e ) = (1 2 5 ln τ ) ( τ + τ ) = (1 2 5 ln τ ) cLx = (1 2 ln τ ) cF x − (1 2 ) . 2 x ln τ = 1 2 x ln τ −2 x ln τ −2 x ln τ −2 x 2x 5.4.2. Integration of the Function y=cF(x) By using the above approach, we obtain the following expression for the in tegral of the function y=cF(x) given by (5.17): ∫ cF ( x ) dx = ∫ ( e( ) + e ( ) ) 5 dx = (1 5 ) ( ∫ e ( ) dx + ∫ e ( ) dx ) 2 x +1 ln τ − 2 x +1 ln τ 2 x +1 ln τ − 2 x +1 ln τ ( 5 ) (1 2 ln τ ) e( ) − (1 2 ln τ ) e ( ) = (1 2 5 ln τ ) ( e ( ) − e ( ) ) = (1 2 5 ln τ ) ( τ −τ ( )) = (1 2 5 ln τ ) sLx = (1 2 ln τ ) sF x + (1 2 ) . 2 x +1 ln τ = 1 2 x +1 ln τ 2 x +1 − 2 x +1 ln τ − 2 x +1 ln τ − 2 x +1 5.4.3. Integration of the Function y=sL(x) ∫ sL ( x ) dx = ∫ ( e ( 2 x +1)ln τ − e −(2 x +1)ln τ ) dx = (1 2 ln τ ) e(2 x +1)ln τ + (1 2 ln τ ) e −(2 x +1)ln τ = (1 2 ln τ ) e(2 x +1)ln τ + e −(2 x +1)ln τ = (1 2 ln τ ) τ2 x +1 + τ −(2 x +1) = ( 5 2 ln τ ) cFx = (1 2 ln τ ) cL x + (1 2) . 5.4.4. Integration of the Function y=cL(x) ∫ cL ( x ) dx = ∫ ( e 2 x ln τ + e −2 x ln τ ) dx = (1 2 ln τ ) e2 x ln τ − (1 2 ln τ ) e −2 x ln τ = (1 2 ln τ ) e2 x ln τ − e −2 x ln τ = (1 2 ln τ ) τ2 x − τ −2 x = ( 5 2 ln τ ) sFx = (1 2 ln τ ) sL x − (1 2) . Alexey Stakhov THE MATHEMATICS OF HARMONY 270 The results of integration are presented in Table 5.1. Table 5.1. Formulas for the integrals of the hyperbolic Fibonacci and Lucas functions sF ( x ) y 1 ∫ ydx cL ( x ) = 5 ln τ 2 1 cF 2 ln τ ( ) x − cF ( x ) 1 2 1 5 ln τ 2 sL ( x ) = sL ( x ) y 5 ∫ ydx 2 ln τ cF ( x ) = 1 sF 2 ln τ ( ) x + 1 2 cL ( x ) 1 2 ln τ ( ) cL x + 1 5 2 2 ln τ sF ( x ) = 1 2 ln τ ( ) sL x − 1 2 5.4.5. Differentiation of the Function y=sF(x) By using the expression (5.25), we obtain the following expressions for the derivatives of the function y=sF(x): y′ = ( 2 ln τ e2 x ln τ + 2 ln τ e −2 x ln τ ) ( = 2 ln τ ) ( 5 ( τ2 x + τ −2 x ) = 2 ln τ ( 5 ) 5 cL ( x ) ; ) y″ = 2 ln τ 5 ( 2 ln τ e2 x ln τ − 2 ln τ e −2 x ln τ ) = ( 2 ln τ ) 5 ( τ2 x − τ −2 x ) = ( 2 ln τ ) sF ( x ) ; 2 2 ( 2 y′′′ = ( 2 ln τ ) = ( 2 ln τ ) 3 ( 5 2 ln τ e2 x ln τ + 2 ln τ e −2 x ln τ 5 τ 2 x + τ −2 x ) ) 3 2 = ( 2 ln τ ) 5 cL ( x ) = ( 2 ln τ ) y′; ... 2 k +1 2 ln τ ) ( 2k (2 k +1) y = cL ( x ) = ( 2 ln τ ) y′; 5 2k (2k ) y = ( 2 ln τ ) sF ( x ) . 5.4.6. Differentiation of the Function y=cF(x) By representing the function y=cF(x) in the form y = cF ( x ) = τ( = e( 2 x +1) ln τ 2 x +1) +e ( +τ ( − 2 x +1) − 2 x +1) ln τ 5 5 (5.26) Chapter 5 Hyperbolic Fibonacci and Lucas Functions 271 and repeatedly carrying out the differentiation of (5.26), we obtain the following expressions for the derivatives of the function y=cF(x): ( ) 5 sL ( x ) ; y′ = 2 ln τ y'' = ( 2 ln τ ) cF ( x ) ; 2 3 2 y''' = ( 2 ln τ ) 5 sL ( x ) = ( 2 ln τ ) y′; 4 IV y ( ) = ( 2 ln τ ) cF ( x ) ; ... 2 k +1 2k y(2 k +1) = ( 2 ln τ ) 5 sL( x ) = ( 2 ln τ ) y′; 2k 2k y ( ) = ( 2 ln τ ) cF ( x ) . 5.4.7. Differentiation of the Function y=sLx By representing the function y=sLx given by (5.18) in the form y = sL ( x ) = τ2 x +1 − τ−(2 x +1) = e( 2 x +1) ln τ −e ( − 2 x +1) ln τ , (5.27) and repeatedly carrying out the differentiation of (5.27), we obtain the following expressions for the derivatives of the function y=sLx: y′ = 2 5 ln τ cF ( x ) ; y″ = (2 ln τ )2 sL ( x ) ; y′′′ = (2 ln τ )3 5cF ( x ) = ( 2 ln τ ) y′; 2 IV y ( ) = (2 ln τ )4 sL ( x ) ; y ( ) = (2 ln τ )5 5cF ( x ) = ( 2 ln τ ) y′; ... 4 V y( 2 k +1) = 5 ( 2 ln τ ) 2 k +1 cF ( x ) = ( 2 ln τ ) y′; 2k 2k 2k y ( ) = ( 2 ln τ ) sL ( x ) . 5.4.8. Differentiation of the Function y=cLx By representing the function y=cLx in the form (5.28) y = cL ( x ) = τ2 x + τ−2 x = e 2 x ln τ + e −2 x ln τ and repeatedly carrying out the differentiation of (5.28), we obtain the following expressions for the derivatives of the function y=cLx: y′ = 2 5 ln τ sF ( x ) ; y″ = ( 2 ln τ ) cL ( x ) ; 2 Alexey Stakhov THE MATHEMATICS OF HARMONY 272 y = 5 ( 2 ln τ ) sF ( x ) = ( 2 ln τ ) y′; 3 2 y ( IV ) = ( 2 ln τ ) cL ( x ) ; 4 y(V ) = 5 ( 2 ln τ ) sF ( x ) = ( 2 ln τ ) y′; ... 5 4 y ( 2k +1) = 5 ( 2 ln τ ) 2 k +1 sF ( x ) = ( 2 ln τ ) y′; 2k y (2 k ) = ( 2 ln τ ) cL ( x ) . 2k The results of differentiation are given in Table 5.2. Table 5.2. Formulas for the derivatives of hyperbolic Fibonacci and Lucas functions y y′ sF ( x ) 2 ln τ cL ( x ) 5 y y 2 ln τ sL ( x ) sL ( x ) 2 5 ln τcF ( x ) cL ( x ) 2 5 ln τsF ( x ) 5 y ′′ ( 2 ln τ )2 sF ( x ) ( 2 ln τ )2 cF ( x ) ( 2 ln τ )2 sL ( x ) ( 2 ln τ )2 cL ( x ) y ′′′ ( 2 ln τ )2 y ′ ( 2 ln τ )2 y ′ ( 2 ln τ )2 y ′ ( 2 ln τ )2 y ′ ( IV ) ( 2 ln τ )4 sF ( x ) ( 2 ln τ ) 4 cF ( x ) ( 2 ln τ ) 4 sL ( x ) ( 2 ln τ ) 4 cL ( x ) (V ) ( 2 ln τ ) 4 y ′ ( 2 ln τ ) 4 y ′ ( 2 ln τ )4 y ′ ( 2 ln τ )4 y ′ # # # # # y y cF ( x ) (2k ) ( 2 k +1 ) ( 2 ln τ )2 k sF ( x ) ( 2 ln τ )2 k cF ( x ) ( 2 ln τ )2 k sL ( x ) ( 2 ln τ )2 k cL ( x ) ( 2 ln τ )2 k y ′ ( 2 ln τ )2 k y ′ ( 2 ln τ )2 k y ′ ( 2 ln τ )2 k y ′ 5.4.9. The Main Identities for the Hyperbolic Fibonacci and Lucas Functions The hyperbolic Fibonacci and Lucas functions (5.16)(5.19) are a generaliza tion of the classical Fibonacci and Lucas numbers given by Binet formulas (5.12) (5.15). These hyperbolic Fibonacci and Lucas functions (5.16)(5.19) are connected with the classical Fibonacci and Lucas numbers through simple correlations (5.20). Using a geometric representation of the hyperbolic Fibonacci and Lucas functions, we result in a very simple geometric interpretation of the correlations (5.20). The classical Fibonacci and Lucas numbers are as if inscribed into the graphs of the hyperbolic Fibonacci and Lucas functions at the discrete points of the variable x=0,±1,±2,±3,… .Thus, the hyperbolic Fibonacci and Lucas functions are an ex tension of the classical Fibonacci and Lucas numbers for the continuous domain. From this it appears that the hyperbolic Fibonacci and Lucas functions possess recursive properties similar to those of the classical Fibonacci and Lucas numbers. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 273 On the other hand, functions (5.16)(5.19) are similar to the classical hyper bolic functions (5.7) and (5.8). We can assume by analogy that the hyperbolic Fibonacci and Lucas functions possess hyperbolic properties similar to the prop erties of classical hyperbolic functions. Let us examine these analogies in greater detail, starting with the sim plest recursive relation that gives classical Fibonacci numbers: Fn+1= Fn + Fn1. (5.29) If we accept n=2k or n=2k+1, where k=0,±1,±2,±3,…, we can rewrite the recursive relation (5.29) in two ways: F2k+1= F2k + F2k1 ( 5.30) F2k+2=F2k+1+F 2k. (5.31) Using the correlations (5.20), we can rewrite the recursive relations (5.30) and (5.31) in terms of hyperbolic Fibonacci and Lucas functions as follows: cF(k)=sF(k)+cF(k1), (5.32) sF(k+1)=cF(k)+sF(k), (5.33) where k is a discrete variable that takes its values from the set {0,±1,±2,±3,…}. Substituting the discrete variable k with the continuous variable x in the recursive relations (5.32) and (5.33), we obtain two important identities for the hyperbolic Fibonacci and Lucas functions: cF(x)=sF(x)+cF(x1) (5.34) sF(x+1)=cF(x)+sF(x). (5.35) Let us represent the recursive relation Ln+1= Ln + Ln1 for the classical Lucas numbers in the form of two recursive relations L2k+1=L2k +L2k1 (5.36) L2k+2=L2k+1+L2k, (5.37) where k is a discrete variable that takes its values from the set {0,±1,±2,±3,…}. Then we represent (5.36) and (5.37) in terms of the hyperbolic Fibonacci and Lucas functions, rewriting the recursive relations (5.36) and (5.37) as follows: sL(k)=cL(k)+sL(k1) (5.38) cL(k+1)=sL(k)+cL(k). (5.39) Substituting the continuous variable x for the discrete variable k in the re cursive relations (5.38) and (5.39), we obtain two important identities: sL(x)=cL(x)+sL(x1) (5.40) cL(x+1)=sL(x)+cL(x). (5.41) Alexey Stakhov THE MATHEMATICS OF HARMONY 274 Note that there is another way to obtain the identities (5.34), (5.35), (5.40), and (5.41). For example, let us prove the identity (5.35) by using the definitions (5.16) and (5.17): sF ( x ) + cF ( x ) = ( τ2 x − τ −2 x ) = τ2 x ( τ + 1) + τ −2 x ( τ −1 − 1) 5 + τ2 x +1 + τ −(2 x +1) 5 5 = τ2( x +1) − τ −2( x +1) 5 = sF ( x + 1) . By analogy we can prove the other identities (5.34), (5.40) and (5.41). The above examination has very unusual consequences for the Fibonacci number theory [13, 16, 28]. We can assume the following hypothesis: all well known identities for the Fibonacci and Lucas numbers have the “hyperbolic in terpretations” in the form of corresponding identities for the hyperbolic Fibonacci and Lucas functions. For example, let us give the “hyperbolic interpretation” of the famous Cassini formula that connects three adjacent Fibonacci numbers: (5.42) Fn2 − Fn −1Fn +1 = ( −1) , where n=0,±1,±2,±3,… is a discrete variable. We can represent the Cassini formula (5.42) in the form of two formulas for the even (n=2k) and odd (n=2k+1) values of the discrete variable n as follows: n +1 F22k − F2 k −1F2 k +1 = −1 (5.43) (5.44) F22k +1 − F2 k F2 k + 2 = 1, where k is a discrete variable that takes its values from the set {0,±1,±2,±3,…}. By using (5.20), we can represent the identities (5.43) and (5.44) as follows: sF 2 ( k ) − cF ( k − 1) cF ( k ) = −1 (5.45) sF 2 ( x ) − cF ( x − 1) cF ( x ) = −1 (5.47) (5.46) cF 2 ( k ) − sF ( k ) sF ( k + 1) = 1. Substituting the continuous variable x for the discrete variable k in the for mulas (5.45) and (5.46), we obtain two important identities for the hyperbolic Fibonacci and Lucas functions: (5.48) cF 2 ( x ) − sF ( x ) sF ( x + 1) = 1. We can prove the identities (5.47) and (5.48) by using the definitions (5.16) and (5.17). For example, let us prove the identity (5.48): (( τ ( 2 x +1 + τ −(2 x +1) ) / 5 ) − (( τ − τ ) / 5 ) × (( τ 2 2x −2 x ) ( 2 x +2 − τ −(2 x +2) ) / 5 ) = ( τ4 x +2 + 2τ2 x +1τ −(2 x +1) + τ −(4 x +2) ) / 5 − ( τ4 x +2 − τ2 − τ −2 + τ −(4 x +2) ) / 5 = ( 2 + τ2 + τ −2 ) / 5. ) (5.49) Chapter 5 Hyperbolic Fibonacci and Lucas Functions 275 In Chapter 2 we derived the Binet formulas for Lucas numbers given by (2.67). For the even n=2k the formula (2.67) can be written as follows: L2k=τ2k+τ2k. For k=1 we have: (5.50) L 2=τ 2+τ 2=3. (5.51) Substituting (5.51) into (5.49), we obtain the identity (5.48). It is important to note that the hyperbolic Fibonacci and Lucas functions are a generalization and extension of the Fibonacci and Lucas numbers for the continuous domain. This means that the classical Fibonacci number theory [13, 16, 28] is reducible to a more general theory the theory of the hyperbolic Fi bonacci and Lucas functions. All identities for the classical Fibonacci and Lucas numbers have their continuous analogs in the form of the corresponding identi ties for the hyperbolic Fibonacci and Lucas functions, and conversely. As practical examples for students we can consider the following “recursive” identities for the hyperbolic Fibonacci and Lucas functions: 1 sF ( x ) cF ( x ) = sL ( 2 x ) − 1 5 1 sF ( x ) sL ( x ) = 5sF ( x ) sF x + 2 1 cF ( x ) cL ( x ) = 5 cF ( x ) cF x − 2 1 1 sF ( x ) = sL x − 2 5 1 1 cF x − = cL ( x ) 2 5 2y −1 2x + 1 + sL2 cF ( x ) + cF ( y ) = 5 cL2 4 4 x sF ( 3 x ) = sF cL ( x ) + 1 2 x cL ( 3 x ) = sF cL ( x ) + 1 2 Let us next consider some “hyperbolic” properties of the hyperbolic Fi bonacci and Lucas functions. Consider these important properties for the classical hyperbolic functions: Alexey Stakhov THE MATHEMATICS OF HARMONY 276 Evenness property sh ( − x ) = − sh ( x ) ; ch ( − x ) = ch ( x ) ; th ( − x ) = th ( x ) . Formulas for addition sh ( x + y ) = sh ( x ) ch ( y ) + sh ( y ) ch ( x ) ; ch ( x + y ) = ch ( x ) ch ( y ) + sh ( y ) sh ( x ) . Formulas for the double angle sh ( 2 x ) = 2ch ( x ) sh ( x ) ; ch ( 2 x ) = 2ch 2 ( x ) − 1 It is easy to prove that the hyperbolic Fibonacci and Lucas functions possess similar “hyperbolic” properties, for example, sF ( 2 x ) = sF ( x ) cL ( x ) cF ( 2 x + 1) = cF ( x ) sL ( x ) sF ( x ) sL ( x ) = sF ( 2 x ) − 1 cF ( x ) cL ( x ) = cF ( 2 x ) + 1 sL ( 2 x ) = sL ( x ) cL ( x ) + 1 1 sF ( y ) sF ( x ) = sL ( x + y ) − sL ( x − y ) 5 sL ( x ) cL ( x ) = sL ( x + y ) + sL ( x − y ) x+y x−y sF ( x ) + sF ( y ) = sF cL 2 2 x−y x+y sF ( x ) − sF ( y ) = sF cL 2 2 x+y x−y cF ( x ) + cF ( y ) = cF cL 2 2 x−y x+y cF ( x ) − cF ( y ) = sF sL 2 2 ctF ( 2 x ) = sF 2 ( x ) + cL2 ( x ) 2 sF ( x ) cL ( x ) The above formulas produce only a small portion of the huge number of math ematical identities for the hyperbolic Fibonacci and Lucas functions. Their proof can become a fascinating pastime for students interested in Fibonacci mathe matics and its applications. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 277 5.5. Symmetric Hyperbolic Fibonacci and Lucas Functions (Stakhov Rozin Definition) The above hyperbolic Fibonacci and Lucas functions given by (5.16)(5.19) have an essential shortcoming relative to the classical hyperbolic functions. In par ticular, they don’t possess the evenness property of classical hyperbolic functions: sh ( − x ) = − sh ( x ) ; ch ( − x ) = ch ( x ) ; th ( − x ) = th ( x ) . (5.52) To overcome this shortcoming, Stakhov and Rozin introduced [106] the so called Symmetric Hyperbolic Fibonacci and Lucas Functions: Symmetric hyperbolic Fibonacci sine sFs ( x ) = τ x − τ− x cFs ( x ) = τ x + τ− x 5 Symmetric hyperbolic Fibonacci cosine 5 Symmetric hyperbolic Lucas sine sLs ( x ) = τ x − τ− x Symmetric hyperbolic Lucas cosine (5.53) (5.54) (5.55) (5.56) cLs ( x ) = τ x + τ− x The Fibonacci and Lucas numbers are determined identically by the sym metric hyperbolic Fibonacci and Lucas functions as follows: sFs ( n ) , for n = 2k Fn = cFs ( n ) , for n = 2k + 1 cLs ( n ) , for n = 2k . Ln = sLs ( n ) , for n = 2k + 1 (5.57) It is easy to prove that the function (5.53) is an odd function because sFs ( − x ) = τ− x − τ x 5 =− τ x − τ− x 5 = − sFs ( x ) . (5.58) On the other hand, cFs ( − x ) = τ− x + τ x 5 = cFs ( x ) , (5.59) Alexey Stakhov THE MATHEMATICS OF HARMONY 278 that is, the symmetric hyperbolic Fibonacci cosine (5.54) is an even func tion. By analogy we can prove the following properties of the hyperbolic Lucas functions: sL ( − x ) = − sL ( x ) (5.60) cL ( − x ) = cL ( x ) . (5.61) The properties (5.58)(5.61) show that the symmetric hyperbolic Fibonacci and Lucas functions (5.52)(5.55) possess the evenness property (5.52). It is easy to construct the graphs of symmetric hyperbolic Fibonacci and Lucas sines and cosines (Fig. 5.3 and Fig. 5.4). Their graphs have symmetric form and in this respect are similar to the classical hyperbolic functions. Y Y y = cLs(x) y = cFs(x) 1 1 y = sFs(x) 0 1 Figure 5.3. A graph of the symmetric hyperbolic Fibonacci sine and cosine y = sLs(x) 0 1 Figure 5.4. A graph of the symmetric hyperbolic Lucas sine and cosine Here it is necessary to point out that at x=0 the symmetric hyperbolic Fibonac ci cosine cFs(x) takes the value cFs(0) = 2 5 , and the symmetric hyperbolic Lucas cosine cLs(x) takes the value cLs(0)=2. It is also important to emphasize that the Fibonacci numbers Fn with even indices (n=0,±2,±4,±6,…) are “inscribed” into the graph of symmetric hyperbolic Fibonacci sine sFs(x) at the discrete points (x=0,±2,±4,±6,…) and the Fibonacci numbers with odd indices (n=±1,±3,±5,…) are “inscribed” into the symmetric hyperbolic Fibonacci cosine cFs(x) at the dis crete points (x=±1,±3,±5,…). On the other hand, the Lucas numbers Ln with the even indices are “inscribed” into the graph of symmetric hyperbolic Lucas cosine cLs(x) at the discrete points (n=0,±2,±4,±6,…) and the Lucas numbers with odd indices are “inscribed” into the graph of the symmetric hyperbolic Lucas cosine sLs(x) at the discrete points (x=±1,±3,±5,…). The symmetric hyperbolic Fibonacci and Lucas functions are connected amongst themselves by the following simple correlations: Chapter 5 sFs ( x ) = Hyperbolic Fibonacci and Lucas Functions 279 sLs ( x ) ; cFs ( x ) = cLs ( x ) . 5 5 Also we can introduce the notions of symmetric hyperbolic Fibonacci and Lucas tangents and cotangents. Symmetric hyperbolic Fibonacci tangent tFs ( x ) = sFs ( x ) cFs ( x ) = τ x − τ− x τ x + τ− x (5.62) Symmetric hyperbolic Fibonacci cotangent ctFs ( x ) = cFs ( x ) sFs ( x ) = τ x + τ− x τ x − τ− x (5.63) Symmetric hyperbolic Lucas tangent tLs ( x ) = sLs ( x ) cLs ( x ) = τ x − τ− x τ x + τ− x (5.64) Symmetric hyperbolic Lucas cotangent ctLs ( x ) = cLs ( x ) sLs ( x ) = τ x + τ− x τ x − τ− x (5.65) We conclude from a comparison of the functions (5.62) with (5.63), and (5.64) with (5.65) that these functions coincide, that is, we have: tFs ( x ) = tLs ( x ) and ctFs ( x ) = ctLs ( x ) . (5.66) It is easy to prove that the functions (5.62) and (5.63) are odd functions because τ− x − τ x τ x − τ− x tFs ( − x ) = − x = − = −tFs ( x ) τ + τx τ x + τ− x τ− x + τ x τ x + τ− x ctFs ( − x ) = − x = − = −ctFs ( x ) . τ − τx τ x − τ− x Taking into consideration (5.66), we can write: tLs ( − x ) = − tLs ( x ) ; ctLs ( − x ) = −ctLs ( x ) , that is, the functions (5.64) and (5.65) are also odd functions. Next let us compare the classical hyperbolic functions (5.7), (5.8) with the symmetric hyperbolic Fibonacci functions (5.53)(5.56). It follows from this comparison that the symmetric hyperbolic Fibonacci and Lucas func tions possess all the important properties of classical hyperbolic functions. Thus, the symmetric hyperbolic Fibonacci and Lucas functions have Hy perbolic Properties. Alexey Stakhov THE MATHEMATICS OF HARMONY 280 On the other hand, a comparison of Fibonacci and Lucas numbers with the symmetric hyperbolic Fibonacci and Lucas functions show that according to (5.61) these functions are a generalization of Fibonacci and Lucas numbers for the continuous domain. This means that the symmetric hyperbolic Fibonacci and Lucas functions possess Recursive Properties similar to the properties of Fi bonacci and Lucas numbers. 5.6. Recursive Properties of the Symmetric Hyperbolic Fibonacci and Lucas Functions Consider the recursive properties of the symmetric hyperbolic Fibonacci and Lucas functions in comparison with the analogous properties of Fibonac ci and Lucas numbers. Theorem 5.1. The following correlations, that are analogous to the recursive correlation for the Fibonacci numbers Fn+2=Fn+1+Fn, are valid for the symmetric hyperbolic Fibonacci functions: sFs ( x + 2 ) = cFs ( x + 1) + sFs ( x ) (5.67) cFs ( x + 2 ) = sFs ( x + 1) + cFs ( x ) . Proof: cFs ( x + 1) + sFs ( x ) = = = − x +1) τ x ( τ + 1) − τ − x (1 − τ ) 5 τ x +2 −( x +2) −τ 5 5 −( x +2) τ x − τ− x 5 τ x × τ 2 − τ − x × τ −2 5 τ x +1 − τ ( 5 τ x ( τ + 1) + τ − x (1 − τ ) x +2 = + = sFs ( x + 2) ; sFs ( x + 1) + cFs ( x ) = = τ x +1 + τ ( 5 − x +1) = + τ x + τ− x 5 τ x × τ 2 + τ − x × τ −2 5 +τ = cFs ( x + 2) . 5 Theorem 5.2. The following correlations, that are analogous to the recursive equation for the Lucas numbers Ln+2=Ln+1+Ln, are valid for the symmetric hy perbolic Lucas functions: = τ Chapter 5 Hyperbolic Fibonacci and Lucas Functions 281 sLs ( x + 2 ) = cLs ( x + 1) + sLs ( x ) (5.68) cLs ( x + 2 ) = sLs ( x + 1) + cLs ( x ) . The proof is analogous to Theorem 5.28. Theorem 5.3 (a generalization of Cassini’s formula). The following corre n +1 lations, that are similar to Cassini’s formula Fn2 − Fn +1Fn −1 = ( −1) , are valid for the symmetric hyperbolic Fibonacci functions: sFs ( x ) − cFs ( x + 1) cFs ( x − 1) = −1 2 (5.69) cFs ( x ) − sFs ( x + 1) sFs ( x − 1) = 1. Proof: 2 sFs ( x ) − cFs ( x + 1) cFs ( x − 1) 2 τ x − τ − x τ x +1 + τ −( x +1) τ x −1 + τ−( x −1) = × − 5 5 5 2 = τ2 x − 2 + τ −2 x − ( τ2 x + τ2 + τ −2 + τ −2 x ) 5 = −1; cFs ( x ) − sFs ( x + 1) sFs ( x − 1) 2 τ x + τ − x τ x +1 − τ −( x +1) τ x −1 − τ −( x −1) = × − 5 5 5 2 = ( τ2 x + 2 + τ−2 x − τ2 x − τ2 − τ−2 + τ−2 x 5 ) = 1. Note that for the proof we used the Binet formula for the Lucas numbers: τ2 + τ−2 = L ( 3 ) = 3. Theorem 5.4. The following correlations, that are similar to the identity n L2n − 2 ( −1) = L2 n , are valid for the symmetric hyperbolic Lucas functions: sLs ( x ) + 2 = cLs ( 2 x ) ; cLs ( x ) − 2 = sLs ( 2 x ) . (5.70) The proof is analogous to Theorem 5.3. Theorem 5.5. The following correlations, that are similar to the iden tity F n+1+F n1=Ln, are valid for the symmetric hyperbolic Fibonacci and Lucas functions: 2 cFs ( x + 1) + cFs ( x − 1) = cLs ( x ) sFs ( x + 1) + sFs ( x − 1) = sLs ( x ) . 2 (5.71) Alexey Stakhov THE MATHEMATICS OF HARMONY 282 The proof is analogous to Theorem 5.1. Theorem 5.6. The following correlations, that are similar to the identity Fn+L n=2Fn+1, are valid for the symmetric hyperbolic Fibonacci and Lucas functions: cFs ( x ) + sLs ( x ) = 2sFs ( x + 1) (5.72) sFs ( x ) + cLs ( x ) = 2cFs ( x ) . The proof is analogous to Theorem 5.1. Based on the definitions (5.53) (5.56), we can prove different identities for the symmetric hyperbolic Fibonacci and Lucas functions. We can see some of these identities in Table 5.3. Table 5.3. The identities for Fibonacci and Lucas numbers and for symmetric hyperbolic Fibonacci and Lucas functions Fn + 2 = Fn +1 + Fn Ln + 2 = Ln +1 + Ln Fn = ( −1) n +1 F− n sFs ( x + 2 ) = cFs ( x + 1) + sFs ( x ) cFs ( x + 2 ) = sFs ( x + 1) + cFs ( x ) sLs ( x + 2 ) = cLs ( x + 1) + sLs ( x ) cLs ( x + 2 ) = sLs ( x + 1) + cLs ( x ) sFs ( x ) = − sFs ( − x ) cFs ( x ) = cFs ( − x ) sLs ( x ) = sLs ( − x ) cLs ( x ) = − cLs ( − x ) sFs ( x + 3 ) + cFs ( x ) = 2cFs ( x + 2 ) cFs ( x + 3 ) + sFs ( x ) = 2sFs ( x + 2 ) sFs ( x + 6 ) − cFs ( x ) = 4 sFs ( x + 3 ) cFs ( x + 6 ) − sFs ( x ) = 4cFs ( x + 3 ) [ sFs ( x )] − cFs ( x + 1) cFs ( x − 1) = −1 [cFs ( x )]2 − sFs ( x + 1) sFs ( x − 1) = 1 F2 n +1 = Fn2+1 + Fn2 cFs ( 2 x + 1) = [cFs ( x + 1)] + [ sFs ( x )] 2 sFs ( 2 x + 1) = [ sFs ( x + 1)] + [cFs ( x )] L2n − 2 ( −1) = L2 n Ln + 5 Fn = 2 Ln +1 [ sLs ( x )]2 + 2 = cLs ( 2 x ) sLs ( x ) + cLs ( x + 3 ) = 2 sLs ( x + 2 ) 2 sLs ( x + 1) sLs ( x − 1) − [ cLs ( x )] = −5 sFs ( x + 3 ) − 2cFs ( x ) = sLs ( x ) sLs ( x − 1) + sLs ( x + 1) = 5 sFs ( x ) sLs ( x ) + 5cFs ( x ) = 2cLs ( x + 1) [cLs ( x )]2 − 2 = cLs ( 2 x ) cLs ( x ) + sLs ( x + 3 ) = 2cLs ( x + 2 ) 2 cLs ( x + 1) cLs ( x − 1) − [ sLs ( x )] = 5 cFs ( x + 3 ) − 2 sFs ( x ) = cLs ( x ) cLs ( x − 1) + cLs ( x + 1) = 5cFs ( x ) cLs ( x ) + 5sFs ( x ) = 2sLs ( x + 1) L2n +1 + L2n = 5 F2 n +1 [ sLs ( x + 1)]2 + [cLs ( x )]2 = 5cFs ( 2 x + 1) [cLs ( x + 1)]2 + [ sLs ( x )]2 = 5sFs ( 2 x + 1) Ln = ( −1) L− n n Fn + 3 + Fn = 2Fn +2 sFs ( x + 3 ) − cFs ( x ) = 2 sFs ( x + 1) Fn + 3 − Fn = 2Fn +1 Fn +6 − Fn = 4 Fn + 3 Fn2 − Fn +1Fn −1 = ( −1 ) n +1 Ln + Ln + 3 = 2 Ln + 2 Ln +1Ln −1 − L2n = −5 ( −1) Fn + 3 − 2 Fn = Ln Ln −1 + Ln +1 = 5 Fn 2 2 n n cFs ( x + 3 ) − sFs ( x ) = 2cFs ( x + 1) 2 2 We can see from Table 5.3 that two identities for hyperbolic Fibonacci and Lucas functions correspond to one identity for Fibonacci and Lucas numbers. This fact can be explained very easily. This all depends on the evenness of the index n of the Fibonacci and Lucas numbers. For example, consider the sim plest identity Fn+2 = Fn+1+Fn. If n=2k (n is even), we should use the first identi ty sFs(x+2)= cFs(x+1)+sFs(x); in the opposite case (n=2k+1) we should use the other identity cFs(x+2)= sFs(x+1)+cFs(x). Chapter 5 Hyperbolic Fibonacci and Lucas Functions 283 5.7. Hyperbolic Properties of the Symmetric Hyperbolic Fibonacci and Lu cas Functions and Formulas for Their Differentiation and Integration 5.7.1. Hyperbolic Properties The symmetric hyperbolic Fibonacci and Lucas functions possess “hyper bolic” properties similar to classical hyperbolic functions. Consider some of them in contrast to certain properties of classical hyperbolic functions. Theorem 5.7. The following identity, that is similar to the identity [ch(x)]2 [ch(x)]2 =1, is valid for the symmetric hyperbolic Fibonacci function: 2 2 4 cFs ( x ) − sFs ( x ) = . (5.73) 5 Proof: 2 τ x + τ− x τ x − τ− x cFs ( x ) − sFs ( x ) = − 5 5 2 2 2 τ2 x + 2 + τ−2 x − τ2 x + 2 − τ−2 x 4 = . 5 5 By analogy, we can prove the following theorem for symmetric hyperbolic Lucas functions. Theorem 5.8. = cLs ( x ) − sLs ( x ) = 4. (5.74) Theorem 5.9. The following identity, that is similar to the identity ch(x+y)=ch(x)ch(y)+sh(x)sh(y), is valid for the symmetric hyperbolic Fibonacci function: 2 cFs ( x + y ) = cFs ( x ) cFs ( y ) + sFs ( x ) sFs ( y ) . (5.75) 5 Proof: 2 2 τ x + τ− x τ y + τ− y τ x − τ− x τ y − τ− y × + × 5 5 5 5 x+y x−y −x+y − x−y x+y x−y −x+y − x−y τ +τ +τ +τ +τ −τ −τ +τ = 5 cFs ( x ) cFs ( y ) + sFs ( x ) sFs ( y ) = = 2 ( τ x + y + τ− x − y ) 5 = 2 cFs ( x + y ) . 5 Alexey Stakhov THE MATHEMATICS OF HARMONY 284 Theorem 5.10. The identity, that is similar to the identity ch(xy)=ch(x)ch(y) sh(x)sh(y), is valid for the symmetric hyperbolic Fibonacci function: 2 cFs ( x − y ) = cFs ( x ) cFs ( y ) − sFs ( x ) sFs ( y ) . (5.76) 5 By analogy we can prove the following theorems for the symmetric hyper bolic Lucas functions. Theorem 5.11. 2cLs ( x ± y ) = cLs ( x ) cLs ( y ) ± sLs ( x ) sLs ( y ) . (5.77) Theorem 5.12. The following identities, that are similar to the identity ch(2x)=[ch(x)]2+[sh(x)]2, are valid for the symmetric hyperbolic Fibonacci and Lucas functions: 2 2 2 cFs ( 2 x ) = cFs ( x ) + sFs ( x ) (5.78) 5 2cLs ( 2 x ) = cLs ( x ) + sLs ( x ) . 2 2 (5.79) Theorem 5.13. The following identities, that are similar to the identity sh(2x)=2sh(x)ch(x), are valid for the symmetric hyperbolic Fibonacci and Lucas functions: 1 sFs ( 2 x ) = sFs ( x ) cFs ( x ) (5.80) 5 sLs ( 2 x ) = sLs ( x ) cLs ( x ) . (5.81) Theorem 5.14. The following formulas, that are similar to Moivre’s formu las [ch(x)±sh(x)]n =ch(nx)±sh(nx), are valid for the symmetric hyperbolic Fi bonacci and Lucas functions: n −1 n 2 cFs ( x ) ± sFs ( x ) = (5.81) cFs ( nx ) ± sFs ( nx ) 5 cLs ( x ) ± sLs ( x ) = 2n −1 cFs ( nx ) ± sFs ( nx ) . n (5.82) 5.7.2. Formulas for Differentiation and Integration It is easy to prove the following formulas for differentiation and integration of symmetric hyperbolic Fibonacci and Lucas functions. Theorem 5.15 (formulas for differentiation). The following correlations similar to the nth derivatives of the classical hyperbolic functions ch ( x ) (n) sh( x ), = ch( x ), for n = 2k + 1 ; for n = 2k Chapter 5 Hyperbolic Fibonacci and Lucas Functions 285 ch( x ), for n = 2k + 1 = sh( x ), for n = 2k are valid for the derivatives of the symmetric hyperbolic Fibonacci functions: sh ( x ) (n) cFs ( x ) (n) sFs ( x ) (n) n ( ln τ ) sFs( x ), for n = 2k + 1 = n ( ln τ ) cFs( x ), for n = 2k (5.83) ( ln τ )n cFs ( x ), for n = 2k + 1 . = n ln τ ( ), 2 sFs x for n k = ) ( (5.84) Theorem 5.16 (formulas for integration). The following correlations simi lar to the integrals of the classical hyperbolic functions sh( x ), for n = 2k + 1 ∫∫∫ ch( x)dx = ch( x), for n = 2k ; n ch( x ), for n = 2k + 1 = sh( x ), for n = 2k n are valid for the symmetric hyperbolic Fibonacci and Lucas functions: ∫∫∫ sh( x )dx ∫∫∫ cFs ( x )dx ln τ − n sFs ( x ), for n = 2k + 1 ) ( = ( ln τ )− n cFs ( x ), for n = 2k (5.85) ∫∫∫ cLs ( x)dx ( ln τ ) − n sLs ( x ), for n = 2k + 1 = −n ( ln τ ) cLs ( x ), for n = 2k (5.86) ∫∫∫ sFs ( x )dx ( ln τ )− n cFs ( x ), for n = 2k + 1 = −n ( ln τ ) sFs ( x ), for n = 2k (5.87) ∫∫∫ sLs ( x)dx −n ( ln τ ) cLs ( x ), for n = 2k + 1 = . −n ( ln τ ) sLs ( x ), for n = 2k (5.88) n n n n Alexey Stakhov THE MATHEMATICS OF HARMONY 286 5.8. The Golden Shofar 5.8.1. The Quasisine Fibonacci and Lucas Functions Consider Binet formulas for Fibonacci and Lucas numbers represented in the following form: n τn − ( −1) τ− n Fn = (5.89) 5 Ln = τn + (−1)n τ− n , ( (5.90) ) where τ = 1 + 5 2 is the golden mean and n= 0, ±1, ±2, ±3, … . By comparing Binet formula (5.89) and (5.90) with the symmetric hyper bolic Fibonacci and Lucas functions (5.53)(5.56), we can see that the continu ous functions τx and τx in the formulas (5.53)(5.56) correspond to the discrete sequences τn and τn in formulas (5.89) and (5.90). Then we set up in correspon dence to the alternating sequence (1)n in the Binet formulas (5.89) and (5.90) some continuous function, taking the values 1 and 1 at the discrete points x= 0, ±1, ±2, ±3, …. The trigonometric function cos(πx) is the simplest of them. This reasoning is the basis for introducing the new continuous function that is con nected with Fibonacci and Lucas numbers. Definition 5.1. The following continuous function is called the Quasisine Fibonacci Function (QSFF): τ x − cos ( πx ) τ− x . QF ( x ) = (5.91) 5 There is the following correlation between Fibonacci numbers Fn given by (5.89) and the quasisine Fibonacci function given by (5.91): τn − cos ( πn ) τ − n Fn = Q F ( n ) = , (5.92) 5 where n= 0, ±1, ±2, ±3, … . Definition 5.2. The following continuous function is called the Quasisine Lucas Function (QSLF): QL ( x ) = τ x + cos ( πx ) τ− x . (5.93) The graph of the QSFF is a quasisine curve that passes through all points corresponding to the Fibonacci numbers that are given by (5.89) on the coordi nate plane (Fig. 5.5). The symmetric hyperbolic Fibonacci functions (5.53) and (5.54) (Fig. 5.3) are the envelopes of the quasisine Fibonacci function Q(x). Chapter 5 Hyperbolic Fibonacci and Lucas Functions 287 The graph of the QSLF is a quasisine curve that passes through all points corresponding to the Lucas numbers that are given by (5.90) on the coordinate plane (Fig. 5.6). The symmetric hyperbolic Lucas functions (5.55) and (5.56) (Fig. 5.4) are the envelopes of the quasisine Lucas function QL(x). Y Y y = cLs(x) y = cFs(x) 1 1 0 1 y = FF(x) y = sFs(x) Figure 5.5. A graph of the quasisine Fibonacci function 0 1 y = FL(x) y = sLs(x) Figure 5.6. A graph of the quasisine Lucas function 5.8.2. Recursive Properties of the Quasisine Fibonacci and Lucas Functions It is easy to prove the following theorems for the quasisine Fibonacci and Lucas functions. Theorem 5.17. For the quasisine Fibonacci function there is the following correlation similar to the recursive relation for the Fibonacci numbers F n+2=Fn+1+F n: Q F ( x + 2 ) = Q F ( x + 1) + Q F ( x ) . (5.94) Theorem 5.18. For the quasisine Lucas function there is the following cor relation similar to the recursive relation for the Lucas numbers Ln+2=Ln+1+Ln: QL ( x + 2 ) = QL ( x + 1) + QL ( x ) . (5.95) Theorem 5.19. For the quasisine Fibonacci function there is the following n +1 correlation similar to Cassini’s formula Fn2 − Fn +1Fn −1 = ( −1) : QF ( x ) − QF ( x + 1)QF ( x − 1) = − cos ( πx ) . 2 (5.96) By analogy with Theorems 5.17 5.19 we can prove other identities for the quasisine Fibonacci and Lucas functions (Table 5.4). THE MATHEMATICS OF HARMONY Alexey Stakhov 288 Table 5.4. The identities for the quasisine Fibonacci and Lucas functions Fn + 2 = Fn +1 + Fn Q F ( x + 2 ) = Q F ( x + 1) + Q F ( x ) Fn + 3 + Fn = 2 Fn + 2 Q F ( x + 3 ) + Q F ( x ) = 2Q F ( x + 2 ) Fn + 3 − Fn = 2 Fn +1 Q F ( x + 3 ) − Q F ( x ) = 2Q F ( x + 1) Fn + 6 − Fn = 4 Fn + 3 Q F ( x + 6 ) − Q F ( x ) = 4Q F ( x + 3 ) Fn2 − Fn +1 Fn −1 = ( −1) F2 n +1 = F 2 n +1 n +1 F F 2 2 +F L Ln + 2 = Ln +1 + Ln L F Ln + 3 + Ln = 2 Ln + 2 Ln −1 + Ln +1 = 5 Fn Ln + 5 Fn = 2Ln +1 L L L L F L F F F n L F L Fn + 3 − 2 Fn = Ln Ln +1 Ln −1 − L = −5 ( −1) 2 F 2 n 2 n [Q ( x )] − Q ( x + 1) Q ( x − 1) = − cos ( πx ) Q ( 2 x + 1) = [Q ( x + 1)] + [Q ( x )] Q ( x + 2 ) = Q ( x + 1) + Q ( x ) Q ( x + 3 ) + Q ( x ) = 2Q ( x + 2 ) Q ( x − 1) + Q ( x + 1) = 5Q ( x ) Q ( x ) + 5Q ( x ) = 2Q ( x + 1) Q ( x + 3 ) − 2Q ( x ) = Q ( x ) Q ( x + 1) Q ( x − 1) − [Q ( x )] = −5 cos ( πx ) L F L 2 L L L 5.8.3. Threedimensional Fibonacci Spiral It is well known that the trigonometric sine and cosine can be defined as a horizontal projection of the translational movement of a point on the surface of an infinite rotating cylinder with radius 1 and the symmetry center coinciding with the axis ОХ. Such threedimensional spiral is described by the complex function f(x)=cos(x)+isin(x), where i = −1. The sine function is its projection on a plane. If we assume that the quasisine Fibonacci function (5.91) is a projection of the threedimensional spiral on some funnelshaped surface, we can then define the socalled Threedimensional Fibonacci Spiral. Definition 5.3. The following function is called the Threedimensional Fi bonacci Spiral: sin ( πx ) τ− x . +i (5.97) 5 5 This function, by its shape, reminds one of a spiral that is drawn on a well with the end bent (Fig. 5.7). It is easy to prove the following theorem for the threedimensional Fibonacci spiral. Theorem 5.20. For the threedimensional Fibonacci spiral the following cor relation, similar to the recursive relation for Fibonacci numbers Fn+2=Fn+1+Fn, is valid: SF ( x ) = τ x − cos ( πx ) τ− x S F ( x + 2 ) = S F ( x + 1) + S F ( x ) . Chapter 5 Hyperbolic Fibonacci and Lucas Functions 289 Y Z 0 X Figure 5.7. The threedimensional Fibonacci spiral 5.8.4. The Golden Shofar We can separate the real and imaginary parts of the threedimensional Fi bonacci spiral (5.97): Re S F ( x ) = τ x − cos ( πx ) τ− x 5 (5.98) sin ( πx ) τ− x Im S F ( x ) = . (5.99) 5 The following system of equations can be obtained from (5.97), (5.98), and (5.99) if we consider the axis OY as a real axis and the axis OZ as an imaginary axis: cos ( πx ) τ− x τx =− y ( x ) − 5 5 −x sin ( πx ) τ z ( x ) = 5 (5.100) Let us square both expressions of the equation system (5.100) and add them together. Taking y and z as independent variables, we obtain a curvilinear sur face of the second degree called the Golden Shofar. THE MATHEMATICS OF HARMONY Alexey Stakhov 290 Definition 5.4. The following curvilinear function of the second degree is the Golden Shofar: 2 2 τ− x τx 2 + = y − z . (5.101) 5 5 We can see in Fig. 5.8 a threedimension al surface corresponding to (5.101). It is sim ilar to the horn or well with a narrow end. In the Hebrew language the word “Shofar” means horn which is a symbol of power. The formula for the Golden Shofar can be represented in the following form: z 2 = cFs ( x ) − y × sFs ( x ) + y , (5.102) Figure 5.8. The Golden Shofar where sFs(x) and cFs(x) are the symmetric hyperbolic Fibonacci sine and cosine, respectively. A projection of the Golden Shofar on the plane XOY is shown in Fig. 5.9. The Golden Shofar is a projection into the space between the graphs of the symmet ric hyperbolic Fibonacci sine and cosine (Fig. 5.3). Y Z y = cFs(x) z = ax/√5 1 0 1 1 X 0 y = sFs(x) 1 X z = ax/√5 Figure 5.10. A projection of the Golden Shofar on the plane XOZ Figure 5.9. A projection of the Golden Shofar on the plane XOY The function (5.97) lies on the Golden Shofar and “pierces” the plane XOY at the points that correspond to the Fibonacci sequence (Fig. 5.9). A projection of the Golden Shofar on the plane XOZ is shown in Fig. 5.10. The Golden Shofar is a projection into the space between the graphs of the two exponent functions − τ x 5 and τ − x 5 . By cutting the Golden Shofar by the planes that are parallel to the plane YOZ, the circles with the center 0;τ x 5 and the radius τ − x 5 are obtained. ( ) ( ( ) ) Chapter 5 Hyperbolic Fibonacci and Lucas Functions 291 It is possible to say that the function y ( x ) = τ x 5 is the pseudoaxis of sym metry (or the axis of pseudosymmetry) of the Golden Shofar (Fig. 5.8). It is possible to assume that the Golden Shofar is a new model of the field with curvilinear structure and similar to the model of the gravitational well used in the general theory of relativity. 5.8.5. A General Model of the Hyperbolic Space with a “Shofarable” Topology Based on experimental data obtained in 2003 by the NASA Wilkinson Microwave Anisotropy Probe (WMAP), a new hypothesis about the structure of the Universe was developed. According to [155], the geometry of the Uni verse is similar in shape to a horn or a pipe with an extended bell. As a result of this discovery we can make the following claim: The Universe has a “shofarlike” topology as shown in Fig. 5.11. Figure 5.11. The “shofarlike” topology 5.9. A General Theory of the Hyperbolic Functions 5.9.1. A Definition of the Hyperbolic Fibonacci and Lucas mFunctions Alexey Stakhov and Boris Rozin introduced [106] a new class of hyperbol ic functions, the Symmetric Hyperbolic Fibonacci and Lucas Functions, based on an analogy between Binet formulas and the classical hyperbolic functions. By using this approach, Alexey Stakhov introduced [118] the Hyperbolic Fi bonacci and Lucas Functions of the Order m or simply Hyperbolic Fibonacci and Lucas mFunctions. These functions are based on an analogy between Gaza le formulas that are given by (4.281) and (4.292) with the classical hyperbolic Alexey Stakhov THE MATHEMATICS OF HARMONY 292 functions (5.7) and (5.8). Consider the new class of hyperbolic Fibonacci and Lucas functions introduced in [118]. Hyperbolic Fibonacci msine x −x Φ mx − Φ m− x m + 4 + m2 m + 4 + m2 1 − sFm ( x ) = = (5.103) 2 2 4 + m2 4 + m 2 Hyperbolic Fibonacci mcosine x −x 2 m + 4 + m2 Φ mx + Φ m− x 1 4 m + + m + (5.104) cFm ( x ) = = 2 2 2 2 4+m 4 + m Hyperbolic Lucas msine x −x m + 4 + m2 m + 4 + m2 x −x − sLm ( x ) = Φ m − Φ m = (5.105) 2 2 Hyperbolic Lucas mcosine x −x m + 4 + m2 m + 4 + m2 x −x + cLm ( x ) = Φ m + Φ m = (5.106) 2 2 It is easy to prove that the Fibonacci and Lucas mnumbers are determined identically by the hyperbolic Fibonacci and Lucas mfunctions as follows: sFm ( n ) for n = 2k Fm ( n ) = cFm ( n ) for n = 2k + 1 (5.107) cLm ( n ) for n = 2k Lm ( n ) = sLm ( n ) for n = 2k + 1 Graphs of the hyperbolic Fibonacci and Lucas mfunctions are similar to graphs of the classical hyperbolic functions. Here it is important to note that at the point x=0, the hyperbolic Fibonacci mcosine cFm(x) (5.104) takes on the value cFm ( x ) = 2 4 + m2 , and the hyperbolic Lucas mcosine cLm(x) takes on the value cLm(0)=2. It is also important to emphasize that the Fibonac ci mnumbers Fm(n) with the even indices n=0,±2,±4,±6,… are “inscribed” into the graph of the hyperbolic Fibonacci msine sFm(x) at the discrete points x=0,±2,±4,±6,… and the Fibonacci mnumbers Fm(n) with odd indices n=±1,±3,±5,… are “inscribed” into the hyperbolic Fibonacci mcosine cFm(x) at the discrete points x=±1,±3,±5,…. On the other hand, the Lucas mnumbers Lm(n) with even indices are “inscribed” into the graph of the hyperbolic Lucas mcosine cLm(x) at the discrete points x=0,±2,±4,±6,… and the Lucas mnum bers Lm(n) with odd indices are “inscribed” into the graph of the hyperbolic Lu cas msine cLm(x) at the discrete points x=±1,±3,±5,… . Chapter 5 Hyperbolic Fibonacci and Lucas Functions 293 We can also introduce the notions of hyperbolic Fibonacci and Lucas m tangents and mcotangents. Hyperbolic Fibonacci mtangent tFm ( x ) = sFm ( x ) cFm ( x ) = Φ mx − Φ m− x Φ mx + Φ m− x (5.108) Hyperbolic Fibonacci mcotangent cFm ( x ) Φ mx + Φ −mx ctFm ( x ) = = sFm ( x ) Φ mx − Φ −mx (5.109) Hyperbolic Lucas mtangent tLm ( x ) = sLm ( x ) cLm ( x ) = Φ mx − Φ −mx Φ mx + Φ −mx (5.110) Hyperbolic Lucas mcotangent ctLm ( x ) = cLm ( x ) sLm ( x ) = Φ mx + Φ m− x Φ mx − Φ −mx (5.111) By analogy we can introduce other hyperbolic Fibonacci and Lucas mfunc tions, in particular, secant, cosecant, and so on. 5.9.2. General Properties of the Hyperbolic Fibonacci and Lucas m Functions It is easy to prove that the function (5.103) is an odd function because sFm ( − x ) = Φ m− x − Φ mx cFm ( − x ) = Φ m− x + Φ mx =− 4 + m2 On the other hand, Φ mx − Φ m− x 4 + m2 = − sFm ( x ) . (5.112) Φ mx + Φ m− x = cFm ( x ) , (5.113) 4 + m2 4 + m2 that is, the hyperbolic Fibonacci mcosine (5.104) is an even function. By analogy, we can prove that the hyperbolic Lucas msine (5.105) is an odd function and the hyperbolic Lucas mcosine (5.106) is an even function, that is, sLm ( x ) = − sLm ( − x ) = (5.114) cLm ( x ) = cLm ( − x ) . (5.115) The properties (5.112)(5.15) show that the symmetric hyperbolic Fibonac ci and Lucas mfunctions (5.103)(5.106) possess the property of evenness (5.52). THE MATHEMATICS OF HARMONY Alexey Stakhov 294 By making the pairwise comparison of the functions (5.108) with (5.110), and (5.109) with (5.111), we can conclude that the hyperbolic Fibonacci and Lucas mtangents and mcotangents are coincident, respectively, that is, we have: tFm ( x ) = tLm ( x ) and ctFm ( x ) = ctLm ( x ) . (5.116) It is easy to prove that the functions (5.108) and (5.109) are odd func tions because Φ − x − Φ mx tFm ( − x ) = −mx = −tFm ( x ) Φ m + Φ mx ctFm ( − x ) = Φ m− x + Φ mx = −ctFm ( x ) . Φ m− x − Φ mx Taking into consideration (5.116) we can write: tLm ( − x ) = −tLm ( x ) ctLm ( − x ) = −ctLm ( x ) , that is, the functions (5.110) and (5.111) are also odd functions. 5.9.3. Partial Cases of the Hyperbolic Fibonacci and Lucas mFunctions Consider the partial cases of the hyperbolic Fibonacci and Lucas mfunc tions (5.103)(5.106) for different values of the order m. Hyperbolic Fibonacci and Lucas 1functions x −x Φ1x − Φ1− x 1 1 + 5 1 + 5 sF1 ( x ) = = − (5.117) 5 5 2 2 x −x 1 1 + 5 1 + 5 Φ1x + Φ1− x cF1 ( x ) = = + (5.118) 5 5 2 2 x −x 1+ 5 1+ 5 x −x − sL1 ( x ) = Φ1 − Φ1 = (5.119) 2 2 −x x 5 1+ 5 cL1 ( x ) + 2 2 Hyperbolic Fibonacci and Lucas 2functions 1+ = Φ1x + Φ1− x = sF2 ( x ) = cF2 ( x ) = Φ 2x − Φ 2− x 8 Φ 2x + Φ 2− x 8 (5.120) ( ) ( ) (5.121) ( ) ( ) (5.122) = x −x 1 1 2 1 2 + − + 2 2 = x −x 1 1 + 2 + 1 + 2 2 2 Chapter 5 Hyperbolic Fibonacci and Lucas Functions 295 ( ) − (1 + 2 ) cL ( x ) = Φ + Φ = (1 + 2 ) + (1 + 2 ) sL2 ( x ) = Φ 2x − Φ 2− x = 1 + 2 x 2 2 −x 2 x −x x −x (5.123) (5.124) Hyperbolic Fibonacci and Lucas 3functions x −x Φ 3x − Φ 3− x 1 3 + 13 3 + 13 sF3 ( x ) = = − 2 2 13 13 x −x x −x Φ + Φ3 1 3 + 13 3 + 13 cF3 ( x ) = 3 = + 2 2 13 13 sL3 ( x ) 3+ = Φ 3x − Φ 3− x = x 13 3 + 13 − 2 2 x (5.125) (5.126) −x (5.127) −x 13 3 + 13 cL3 ( x ) + (5.128) 2 2 Note that a list of these functions can be continued ad infinitum. It is easy to see that the functions (5.103)(5.106) are connected by very simple correlations: 3+ = Φ 3x + Φ 3− x = sFm ( x ) = sLm ( x ) ; cFm ( x ) = cLm ( x ) . (5.129) 4 + m2 4 + m2 This means that the hyperbolic Lucas mfunctions (5.103) and (5.104) coin cide with the hyperbolic Fibonacci mfunctions (5.105) and (5.106) to within the constant coefficient 1 1 + m2 . 5.9.4. Comparison of the Classical Hyperbolic Functions with the Hyperbolic Lucas mFunctions Let us compare the hyperbolic Lucas mfunctions (5.105) and (5.106) with the classical hyperbolic functions (5.7) and (5.8). For the case 4 + m2 + m =e (5.130) 2 the hyperbolic Lucas mfunctions (5.105) and (5.106) coincide with the classical hyperbolic functions (5.7) and (5.8) to within the constant coefficient 1/2, that is, sL ( x ) cLm ( x ) sh ( x ) = m . and ch ( x ) = (5.131) 2 2 By using (5.130), after simple transformations we can calculate the value Φm = Alexey Stakhov THE MATHEMATICS OF HARMONY 296 me, for which the equality (5.130) is valid: 1 me = e − ≈ 2.35040238... (5.132) e Thus, according to (5.131) the classical hyperbolic functions (5.7) and (5.8) are a partial case of the hyperbolic Lucas mfunctions, if m is equal to (5.132). As the classical hyperbolic functions (5.7) and (5.8) are a partial case of (5.105) and (5.106), we have the right to assert that the formulas (5.103) through (5.106) represent a general class of hyperbolic functions. 5.9.5. Recursive Properties of the Hyperbolic Fibonacci and Lucas m Functions The hyperbolic Fibonacci and Lucas mfunctions possess recursive prop erties similar to Fibonacci and Lucas mnumbers that are given by the recur sive relations (4.250) and (4.302). On the other hand, they possess all of the hyperbolic properties similar to the properties of the classical hyperbolic functions. Let us first prove the recursive properties for the hyperbolic Fi bonacci and Lucas mfunctions. Theorem 5.21. The following correlations that are similar to the recursive relation for the Fibonacci mnumbers Fm(n+2)=mFm(n+1)+Fm(n) are valid for the hyperbolic Fibonacci mfunctions: sFm ( x + 2 ) = mcFm ( x + 1) + sFm ( x ) (5.133) cFm ( x + 2 ) = msFm ( x + 1) + cFm ( x ) . Proof: Φ x +1 + Φ m− x −1 Φ mx − Φ m− x mcFm ( x + 1) + sFm ( x ) = m m + 4 + m2 4 + m2 (5.134) = Φ mx ( mΦ m + 1) − Φ m− x (1 − mΦ m−1 ) . (5.135) 4 + m2 As mΦ m + 1 = Φ 2m and 1 − Φ m−1 = Φ m−2 , we can represent (5.135) as follows: Φ x + 2 − Φ −mx −2 mcFm ( x + 1) + sFm ( x ) = m = sFm ( x + 2 ) 4 + m2 that proves the identity (5.133). By analogy, we can prove the identity (5.134). Theorem 5.22 (a generalization of Cassini formula). The following correlations, that are similar to the Cassini formula n +1 2 Fm ( n ) − Fm ( n − 1) Fm ( n + 1) = ( −1) , are valid for the hyperbolic Fibonacci mfunctions: Chapter 5 Hyperbolic Fibonacci and Lucas Functions 297 sFm ( x ) − cFm ( x + 1) cFm ( x − 1) = −1 2 cFm ( x ) − sFm ( x + 1) sFm ( x − 1) = 1. Proof: 2 (5.136) (5.137) sFm ( x ) − cFm ( x + 1) cFm ( x − 1) 2 2 Φ x − Φ m− x Φ mx+1 + Φ m− x −1 Φ mx−1 + Φ m− x +1 = m × − 4 + m2 4 + m2 4 + m2 = = Φ 2mx − 2 + Φ m−2 x − ( Φ 2mx + Φ 2m + Φ m−2 + Φ m−2 x ) −2 − ( Φ 2m + Φ m−2 ) 4 + m2 (5.138) . 4 + m2 By using the Binet formula (4.292), for the case n=2 we can write: Lm ( 2 ) = Φ 2m + Φ m−2 . (5.139) By using the recursive formula (4.290) and the seeds (4.287) and (4.288), we can represent the Lucas mnumber Lm(2) as follows: (5.140) Lm ( 2 ) = mLm (1) + Lm ( 0 ) = m × m + 2 = m 2 + 2. Taking into consideration (5.140), we can conclude from (5.138) that the identity (5.136) is valid. By analogy, we can prove the identity (5.137). 5.9.6. Hyperbolic Properties of the Hyperbolic Fibonacci and Lucas m Functions Theorem 5.23. The following identity, that is similar to the identity [ch(x)]2 [sh(x)]2=1 for the classical hyperbolic functions, is valid for the hyperbolic Fi bonacci mfunctions: 2 2 4 cFm ( x ) − sFm ( x ) = . (5.141) 4 + m2 Proof: 2 2 2 2 Φ mx + Φ m− x Φ mx − Φ m− x cFm ( x ) − sFm ( x ) = − 4 + m2 4 + m2 Φ 2 x + 2 + Φ m−2 x − Φ 2mx + 2 − Φ m−2 x 4 = m . = 2 4+m 4 + m2 Theorem 5.24. The following identity, that is similar to the identity [ch(x)]2 [sh(x)]2=1 for the classical hyperbolic functions, is valid for the hyper bolic Lucas mfunctions: Alexey Stakhov THE MATHEMATICS OF HARMONY 298 2 2 (5.142) cLm ( x ) − sLm ( x ) = 4. The proof is analogous to Theorem 5.23. Theorem 5.25. The following identity, that is similar to the identity ch(x+y)=ch(x)ch(y)+sh(x)sh(y) for the classical hyperbolic functions, is valid for the hyperbolic Fibonacci mfunctions: 2 cFm ( x + y ) = cFm ( x ) cFm ( y ) + sFm ( x ) sFm ( y ) . (5.143) 4 + m2 Proof: cFm ( x ) cFm ( y ) + sFm ( x ) sFm ( y ) = Φ mx + Φ m− x × Φ my + Φ m− y + Φ mx − Φ m− x × Φ my − Φ m− y 4 + m2 4 + m2 4 + m2 4 + m2 Φ x + y + Φ mx− y + Φ m− x + y + Φ m− x − y + Φ mx+ y − Φ mx− y − Φ m− x + y + Φ m− x − y = m 4 + m2 2 ( Φ mx+ y + Φ m− x − y ) 2 cFm ( x + y ) . = = 4 + m2 × 4 + m2 4 + m2 Theorem 5.26. The following identity, that is similar to the identity ch(x y)=ch(x)ch(y)sh(x)sh(y) for the classical hyperbolic functions, is valid for the hyperbolic Fibonacci mfunctions: 2 cFm ( x − y ) = cFm ( x ) cFm ( y ) − sFm ( x ) sFm ( y ) . (5.144) 4 + m2 The proof is analogous to Theorem 5.25. By analogy, we can prove the following theorems for the hyperbolic Fibonacci and Lucas mfunctions. Theorem 5.27. The following identities, that are similar to the identity ch(2x)=[ch(x)]2+[sh(x)]2 for the classical hyperbolic functions, are valid for the hyperbolic Fibonacci and Lucas mfunctions: 2 2 2 cFm ( 2 x ) = cFm ( x ) + sFm ( x ) (5.145) 5 2cLm ( 2 x ) = cLm ( x ) + sLm ( x ) . 2 2 (5.146) Theorem 5.28. The following identities, that are similar to the identity sh(2x)=2sh(x)ch(x) for the classical hyperbolic functions, are valid for the hy perbolic Fibonacci and Lucas mfunctions: 1 sFm ( 2 x ) = sFm ( x ) cFm ( x ) (5.147) 4 + m2 Chapter 5 Hyperbolic Fibonacci and Lucas Functions 299 sLm ( 2 x ) = sLm ( x ) cLm ( x ) . (5.148) Theorem 5.29. The following formulas, that are similar to Moivre’s formu las [ch(x)±sh(x)]n =ch(nx)±sh(nx) for the classical hyperbolic functions, are valid for the hyperbolic Fibonacci and Lucas mfunctions: n −1 n 2 cFm ( x ) ± sFm ( x ) = cFm ( nx ) ± sFm ( nx ) (5.149) 2 4+m (5.150) cLm ( x ) ± sLm ( x ) = 2n −1 cFm ( nx ) ± sFm ( nx ) . Thus, our research results in a general theory of hyperbolic functions based on the Gazale formulas (4.281) and (4.291). Over several centuries, science, in partic ular, mathematics and theoretical physics, made wide use of classical hyperbolic functions with base e. These functions were used by Lobachevsky in his nonEu clidean geometry and Minkowski in his geometric interpretation of Einstein’s rel ativity theory. More recently Ukrainian mathematicians Stakhov, Tkachenko and Rozin [51, 98, 106, 116, 118, 119] broke monopoly on classical hyperbolic func tions in contemporary mathematics and theoretical physics. It is now clear that the above hyperbolic Fibonacci and Lucas mfunctions based on the Gazale for mulas infinitely extendable to new hyperbolic models of Nature. It is difficult to imagine that the set of new hyperbolic functions has the same cardinality as the set of real numbers because every positive real number m generates its own kind of hyperbolic functions! And all of them possess unique recursive and hyperbolic properties similar to the properties of classical hyperbolic functions and the sym metric hyperbolic Fibonacci and Lucas functions introduced in [106, 116, 119]. n 5.10. A Puzzle of Phyllotaxis 5.10.1. The Phenomenon of Phyllotaxis In Chapter 2 we described the botanical phenomenon known as Phyllotaxis. This phenomenon is inherent in many biological objects. On the surface of the bioorgans, these objects (sprouts of plants and trees, seeds on the disks of sun flower heads and pine cones, etc.) are arranged in clockwise and counterclock wise spirals. There are two types of phyllotaxis. The first type concerns the dispo sition of branches of plants and trees, and the second type concerns the dense packed phyllotaxis objects such as that of a pinecone, pineapple, or the disk of a sunflower head. In the first case, the regularities of phyllotaxis are described by the ratios of adjacent Fibonacci numbers taken through one, that is, Alexey Stakhov THE MATHEMATICS OF HARMONY 300 (5.151) cLm ( x ) ± sLm ( x ) = 2n −1 cFm ( nx ) ± sFm ( nx ) . Each plant and tree possesses its own characteristic ratio taken from (5.151). This ratio is named the Phyllotaxis Order of a given plant or tree. The phyllotaxis orders are different for different plants and trees, for exam ple, for linden, elm, beech and cereals the phyllotaxis order is equal to (2/1); for alder, hazel and grape (3/1); for oak and cherry (5/2); for raspberry, pear, poplar and barberries (8/3); and for almonds (13/5). Note that trees and plants are subject primarily to the laws of Fibonacci phyllotaxis (5.151); in rare cases the plants are subject to Lucas phyllotaxis (Lucas numbers 2, 1, 3, 4, 7, 11, 18, 29, …) or in a few cases phyllotaxis based on the numerical sequence: 5, 2, 7, 9, 16, 25, … that also satisfies the Fibonacci recursive formula. The other type of phyllotaxis is represented in the form of denselypacked botanical objects such as the head of a sunflower, pinecone, pineapple or cactus, etc. For such phyllotaxis objects, it is used usually the number ratios of the left hand and righthand spirals observed on the surface of the phyllotaxis objects. These ratios are equal to the ratios of adjacent Fibonacci numbers, that is, Fn +1 2 3 5 8 13 21 1+ 5 : , , , , , ,... → τ = . (5.152) Fn 1 2 3 5 8 13 2 For example, the head of a sunflower can have the phyllotaxis orders given by Fibonacci ratios: 89/55, 144/89 and even 233/144. n 5.10.2. Dynamic Symmetry When observing phyllotaxis the question arises: how do Fibonacci spirals forming on the surface during growth? This problem is one of the most intrigu ing Puzzles of Phyllotaxis. Its essence consists in the fact that the majority of bioforms change their phyllotaxis orders during their growth. It is known, for example, that sunflower disks that are located on different levels of the same stalk have different phyllotaxis orders; moreover, the greater the age of the disk, the higher its phyllotaxis order tends to be. This means that during the growth of the phyllotaxis object, a natural modification (or increase) of the symmetry relation occurs, and this modification of symmetry obeys the law: 2 3 5 8 13 21 → → → → → → .... (5.153) 1 2 3 5 8 13 The modification of the phyllotaxis orders according to (5.153) is called Dynamic Symmetry [37]. Many scientists who study this puzzle of phyllotax is believe that the phenomenon of dynamic symmetry (5.153) is of fundamen Chapter 5 Hyperbolic Fibonacci and Lucas Functions 301 tal interdisciplinary importance. Recall that, in Vernadsky’s opinion, the prob lem of biological symmetry is the key problem of biology. Thus, the phenomenon of dynamic symmetry (5.153) plays a special role in the geometric problem of phyllotaxis. One may assume that this numerical reg ularity (5.153) reflects some general geometric laws that hide the secret of the dynamic mechanism of phyllotaxis, and uncovering it would be of great impor tance for understanding the phyllotaxis phenomenon in general. A new geometric theory of phyllotaxis was developed recently by the Ukrai nian architect Oleg Bodnar. This original theory is stated in Bodnar’s books [37, 52]. However, in order to understand Bodnar’s research more fully, we have to get a better understanding of the geometric theory of hyperbolic functions [156]. 5.11. A Geometric Theory of the Hyperbolic Functions 5.11.1. Compression and Expansion We start the study of geometric theory of hyperbolic functions from the impor tant geometric transformation of the hyperbolic geometry. This is a geometric trans formation named the Compression to a Straight Line о with the Compression Coeffi cient k (Fig. 5.12). This transformation consists of the following: Every point А of the plane passes into the point А′ that lies on the ray РА perpendicular to о, here the ratio PA′ : PA=k or PA′ = kPA (Fig. 5.12a). If the compression coefficient k>1, then PA′ > PA (Fig. 5.12b); in this case the transformation could be named an Expansion from a Straight Line о. It is clear that the expansion using the coefficient k is equiva lent to the compression with coefficient (1/k). A compression and an expansion possess A′ a number of important properties [156]: A 1. At the compression (expansion) every A straight line passes on into a straight line. A′ 2. At the compression (expansion) all parallels pass on into parallels. 3. At the compression (expansion) the ra tio of the segments lying on one straight o P o P line remain constant. (a) (b) 4. At the compression (expansion), the Figure 5.12. The compression of a point A to a straight line (a) and the areas of all figures change in a constant ra expansion from a straight line (b) tio equal to the compression coefficient k. Alexey Stakhov THE MATHEMATICS OF HARMONY 302 5.11.2. Hyperbola Consider the important geometric curve called a Hyperbola. It is described by the following equality: y=a/x or xy=a. A graph of the hyperbola is represented in Fig. 5.13. (5.154) y b Q a M Q M M1 P1 O2 O P O M2 a Figure 5.13. Hyperbola P x Q1 P2 b Figure 5.14. The axes of a hyperbola It follows from (5.154) and Fig. 5.13 that a graph of the hyperbola consists of two branches that are located for the case a>0 in the first quadrant (x and y are positive) and in the third quadrant (x and y are negative) of the coordinate sys tem. Geometrically, the branches of the hyperbola aim towards the coordinate axes, but they never intersect themselves. This means that the coordinate axes are asymptotes of the hyperbola. Note that the equation xy=a has a simple geometric interpretation: the area of the rectangles MQOP or M′Q′O′P′ that are bounded by the coordinate axes and the straight lines that are drawn through any points M and M′ of the hyperbola in parallel to the coordinate axes (Fig. 5.13) is equal to a, that is, this area does not depend on the choice of the points M and M′. We will name these rectangles MQOP and M′Q′O′P′ with area equal to a the Coordinate Rectangles of the points M and M′. Then we can give the following geometric def inition of the hyperbola [156]: “A hyperbola is the geometric location of the points that lie in the first and third quadrants of the coordinate system with coordinate rectangles that have constant area.” It is easy to prove that the origin of coordinate O is the Centre of Symmetry of the hyperbola, that is, the branches of the hyperbola are symmetric one to other Chapter 5 Hyperbolic Fibonacci and Lucas Functions 303 with respect to the origin of coordinates O. The hyperbola also has the Axes of Sym metry, the bisectors of the coordinate angles аа and bb (Fig. 5.14). The centre of symmetry O and the axes of symmetry аа and bb are frequently called simply the Center and the Axes of the hyperbola; the points A and B, in which the hyperbola is intersected with the axis аа, are called Tops of the hyperbola. Hereinafter, we will use analogies between hyperbola and circle. With this purpose in mind we will introduce, first of all, the concept of the Diameter of a hyperbola; every line segment passing through the centre of the hyperbola and connecting the points of the opposite branches of the hyperbola is called a Diameter of the Hyperbola (it is similar to the diameter of a circle passing through its centre). Let us also introduce the concept of a radius of the hyperbola; a line segment, going from the centre of the hyperbola up to the crossing point with the hyperbola, is called a radius of the hyperbola (that is, the radii of the hyper bola are determined similarly to the radii of the circle). 5.11.3. Geometric Definition of the Hyperbolic Functions The hyperbola in Fig. 5.13 and Fig. 5.14 is the basis for a geometric definition of hyperbolic functions. The geometric theory of Hyperbolic Func tions or Hyperbolic Trigonometric Functions is similar to the theory of tradi tional Circular Trigonometric Functions. In order to emphasize an analogy between hyperbolic and circular trigonometric functions, we will state the theory of the hyperbolic functions in parallel with the theory of circular trigonometric functions. We can choose the axis of symmetry of the hyper bola by the coordinate axes as is shown in Fig. 5.15 and then we use this geometric representation of the hyperbola for the geometric definition of the hyperbolic functions. Y Y M N M N X B O P1 P A X B O P A P1 N1 M1 (a) M1 N1 (b) Figure 5.15. The unit circle (a) and the unit hyperbola (b) Alexey Stakhov THE MATHEMATICS OF HARMONY 304 Let us examine the unit circle (Fig. 5.15a). We can see in Fig. 5.15a a Circular Sector OMA bounded by the radii OM, OA and the arc MA. The number that is equal to the length of the arc АМ or equal to the double area of the sector, bound ed by the radii OM and OA and the arc MA, is called a Radian Angle α between the radii ОА and ОМ of the circle. We can now drop the perpendicular МР to the diameter ОА from the point М of the circle; at the point A we draw a tangent to the circle up to its intersection with the diameter ОМ at the point N. The line segment РМ of the perpendicular is called a Line of Sine, the line segment ОР of the diameter is called a Line of Cosine and the line segment AN is called a Line of Tangent. The lengths of the line segments РМ, ОР and AN are equal respectively to the Sine, Cosine and Tangent of the angle α, that is, y= y= sh x y = thx x shx y = thx y= PM=sht, OP=cht, AN=tht. As is wellknown, the circular trigono metric functions are changing periodical ly with the period 2π. In contrast with this the hyperbolic functions are not periodic. It follows from Fig. 5.15b that the hyper bolic angle t changes from 0 up to ∞. It fol lows from the definition of hyperbolic functions that at the change of the hyper bolic angle from 0 up to ∞, the hyperbolic sine sht is changing from 0 up to ∞, the hy perbolic cosine cht is changing from 1 up ch x PM=sinα, OP=cosα, AN=tgα. Now, let us examine the unit hyperbola (Fig. 5.15b) X2 – Y2 =1. We can see in Fig.20b the Hyperbolic Sector OMA bounded by the hyperbolic radii OM, OA and the hyperbolic arc MA. Then, the number that is equal to the double area of the hyperbolic sector, bounded by these radii OM, OA and the arc MA of the hyperbola, is called the Hyperbolic Angle t between the hyperbolic radii ОА and ОМ. We can now drop the perpendicular МР from the point М of the hyper bola to the diameter ОА, which is a symmetry axis, intersecting the hyperbola at the top A. Next we draw a tangent to the hyperbola to its intersection with the radius ОМ at point N. The line segment РМ of the perpendicular is called a Line of Hyperbolic Sine, the line segment ОР of the axis X is called a Line of Hyper bolic Cosine and the line segment AN is called a Line of Hyperbolic Tangent. The lengths of the line segments РМ, ОР and AN are equal respectively to the Hyper bolic Sine, Hyperbolic Cosine and Hyper y bolic Tangent of the hyperbolic angle t, that is, Figure 5.16. Graphs of the hyperbolic functions Chapter 5 Hyperbolic Fibonacci and Lucas Functions 305 to ∞ and the hyperbolic tangent tht is changing from 0 up to 1. The graphs of these functions are represented in Fig. 5.16. Based upon this geometric approach, it is easy to obtain the basic rela tions for the circular and hyperbolic trigonometric functions. It follows from the similarity of the triangles ОМР and ОNА (Fig. 5.15a) that AN PM = . OA OP However, AN/OA=tgα (because OA=1), and PM/OP=sinα/cosα. Thus, we have: tgα= sinα/cosα. Further, the coordinates of any point М of the circle are equal OP=X, PM=Y. However, then the following important identity follows from the unit circle equation X2+Y2=1: OP2+PM2 = 1 or cos2α+sin2α=1. By dividing both parts of the obtained identity at first by cos2α and then by sin2α we get the following remarkable formulas for the trigonometric functions: 1+tg2α=1/cos2α ctg2α+1=1/sin2α. It follows from a similarity of the triangles ОМР and ОNА (Fig. 5.15b) that AN PM = . OA OP However, AN/OA=tgt (because OA=1), and PM/OP=sht/cht. Thus, we have: tgt= sht/cht. Further, the coordinates of any point М of the hyperbola are equal OP = X , PM = Y . However, then the following important identity follows from the unit hyperbola equation X2Y2=1: OP2PM2 = 1 or ch2t – sh2t=1. By dividing both parts of the obtained identity at first by ch2t and then by sh2t we get the following remarkable formulas for the hyperbolic functions: 1th2t=1/ch2t cth2t1=1/sh2t. Alexey Stakhov THE MATHEMATICS OF HARMONY 306 By using the geometric approach, we can prove many other identities for the trigonometric and hyperbolic functions. 5.11.4. Hyperbolic Rotation Next we will study the hyperbola xy=a. First we make the compres sion of a plane to the axis x with the compression coefficient k. In this case the hyperbola xy=a passes on into the hyperbola xy=ka because the ab scissa x remains without change and the ordinate y is replaced by yk. Then, we make one more compression of a plane to the axis y with coefficient 1/k. Note that the compression with coefficient 1/k is equivalent to the expan sion with coefficient k. After the fulfilment of the compression to axis y with coefficient 1/k it is equivalent to the extension from axis y with the same coefficient k, the hyperbola xy=ka passes on into the hyperbola xy=(ka/k)=a, because the ordinate y of each point for the case of new com pression to the axis y does not vary, and the abscissa x passes on into x/k. Thus, we can see that the sequential compression of the plane to the axis x with the compression coefficient k and then to the axis y with the compres sion coefficient 1/k transforms the hyperbola xy=a into itself. A sequence of these two compressions of the plane to a straight line represents the im portant geometric transformation called Hyperbolic Rotation. The title of the hyperbolic rotation reflects the fact that in such transformation all points of the hyperbola act as though they “glide on a curve,” that is, the hyperbola acts as though it “rotates.” Once again, note that the hyperbolic rotation is the sequential fulfilment of the two geometric transformations, at first the compression of a plane with the coefficient k to the axis x and then the expansion of a plane from the axis y with the same coefficient k. From the above properties of the “compression” and “extension,” the follow ing properties of the hyperbolic rotation result: 1. At the hyperbolic rotation every straight line passes on into a straight line. 2. At the hyperbolic rotation the coordinate axes (the asymptotes of the hyperbola) pass on into themselves. 3. At the hyperbolic rotation parallels pass on into parallels. 4. At the hyperbolic rotation the ratios of all segments, lying on one and the same straight line, remain constant. 5. At the hyperbolic rotation, the areas of all transferred figures remain constant. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 307 It is very important to emphasize that by means of the choice of the appro priate value of the coefficient k, and by means of the hyperbolic rotation we can transfer each point of the hyperbola into any other point of the same hyperbola. In fact, the compression to the axis x with a given coefficient k transfers the point (x,y) of the hyperbola xy=a into the point (x,ky) of the other hyperbola xy=ka; after that the extension of the point (x,ky) from the axis y with the same coeffi cient k transfers the point (x,ky) of the hyperbola xy=ka into the point (x/y,ky) of the initial hyperbola. Thus, a result of the hyperbolic rotation the point (x,y) passes on into the point (x/y,ky) of the initial hyperbola. It follows from here that by means of a suitable hyperbolic rotation we can transfer the point (x,y) of the hyperbola into the point (x1,y1) of the same hyperbola provided we take the compression coefficient k=x/x1. 5.12. Bodnar’s Geometry 5.12.1. Structural numerical Analysis of Phyllotaxis Lattices As previously noted, the Ukrainian architect Oleg Bodnar recently made an attempt to uncover the Puzzle of Phyllotaxis with a new phyllotaxis geometry (“Bodnar’s geometry”) presented in his books [37, 52]. To understand Bodnar’s ge ometry let us study a cedar cone as a characteristic example of a phyllotaxis object (Fig. 5.17a). On the surface of the cedar cone every seed is blocked by the adjacent seeds in three di a b rections. As a result, we can see the picture consisting of three types of spirals equal to the Fi bonacci numbers: 3, 5 and 8. In an effort to simplify the geo metric model of the phyllotax is object in a Fig. 5.17a, b, we d c can represent the phyllotaxis Figure 5.17. Analysis of structurenumerical object in a cylindrical form (Fig. properties of the phyllotaxis lattice Alexey Stakhov THE MATHEMATICS OF HARMONY 308 5.17c). If we cut the surface of the cylinder in Fig. 5.17c by the vertical straight line and then unroll the cylinder on a plane (Fig. 5.17d), we obtain a fragment of the phyllotaxis lattice that is bounded by two parallel straight lines that are traces of the cutting line. We can see that the three groups of parallel straight lines in Fig. 5.17d, namely, the three straight lines 021, 116, 28 with the right hand small declination; the five straight lines 38, 116, 419, 727, 030 with the lefthand declination; and the eight straight lines 024, 327, 630, 125, 425, 7 28, 218, 521 with the righthand abrupt declination, correspond to three types of spirals on the surface of the cylinder in Fig. 5.17c. We use the following method of numbering the lattice nodes in Fig. 5.17d, introducing the following system of coordinates. We use the direct line OO ′ as the abscissa axis and the vertical trace that passes through the point O as the ordinate axis. Taking the ordinate of the point 1 as the length unit, the number that is ascribed to some point of the lattice is then equal to its ordinate. The lattice that is numbered by the indicated method has some characteristic prop erties. Any pair of the points gives a certain direction in the lattice system and, finally, the set of the three parallel directions of the phyllotaxis lattice. We can see that the lattice in Fig. 5.17d consists of triangles. The vertices of the trian gles are indicated by the letters a, b and c. It is clear that the lattice in Fig. 5.17 d consists of the set triangles of the kind {c,b,a}, for example, {0,3,8}, {3,6,11}, {3,8,11}, {6,11,14} and so on. It is important to note that the sides of the triangle {c,b,a} are equal to the differences between the values a,b,c of the triangle {a,b,c} and are the adjacent Fibonacci numbers: 3,5,8. For example, for the triangle {0,3,8} we have the following differences: 30=3; 83=5; 80=8. This means that the sides of the triangle {0,3,8} are equal to, respectively, 3, 5, 8. For the triangle {3,6,11} we have: 63=3; 116=5; 113=8. This means that its sides are equal to 3, 5, 8, respectively. Here each side of the triangle defines one of three declinations of the straight lines that build up the lattice in Fig. 5.17d. In particular, the side of length 3 defines the righthand small declination, the side of length 5 defines the lefthand declination, and the side of length 8 defines the righthand abrupt dec lination. Thus, Fibonacci numbers 3, 5, 8 determine the structure of the phyllo taxis lattice in Fig. 5.17d. The second property of the lattice in Fig. 5.17d is the following. The line seg ment OO′ can be considered as a diagonal of the parallelogram constructed on the basis of the straight lines corresponding to the lefthand declination and the right hand small declination. Thus, the given parallelogram allows one to evaluate the symmetry of the lattice without having to use digital numbering. This parallelo gram is called a Coordinate Parallelogram. Note that coordinate parallelograms of different sizes correspond to lattices with different symmetries. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 309 5.12.2. The Dynamic Symmetry of Phyllotaxis Objects Here we begin an analysis of the phenomenon of dynamic symmetry. The idea of analysis involves the comparison of a series of phyllotaxis lattices (the unrolling of the cylindrical lattice) with different symmetries (Fig. 5.18). In Fig. 5.18 a variance of Fibonacci phyllotaxis is illustrated, where we ob serve the following modification of the dynamic symmetry of the phyllotaxis object during its growth: 1:2:1→ 2:3:1→ 2:5:3 → 3:8:3 → 5:13:8 Note that the lattices, represented in Fig. 5.18, are con sidered as the sequential stages (5 stages) of the transforma tion of one and the same phyllotaxis object. There is a ques tion: how are the transformations of the lattices being car ried out, that is, which geometric movement can be used to provide the sequential passing of all illustrated stages of the phyllotaxis lattice? 5.12.3. The Key Idea of Bodnar’s Geometry Here we will not go deeply into Bodnar’s original reasoning that resulted in a new geometrical theory of phyllotaxis, but direct readers to the remarkable Bod nar books [37, 52] for a more detailed acquain tance with his original geometry. Rather we focus our attention solely on two key ideas that underlie his geometry. We first begin with an analysis of the phe nomenon of dynamic symmetry. The idea of analysis involves a comparison of the series of the phyllotaxis lattices of different symmetries (Fig. 5.18). We start with the comparison of stages I and II. At these stages the lattice can be transformed by the compres sion of the plane along the 03 direction up to the position Figure 5.18. An analysis of dynamic symmetry of a where the line segment 03 ar rives at the edge of the lattice. phyllotaxis object Alexey Stakhov THE MATHEMATICS OF HARMONY 310 Simultaneously, the expansion of the plane should occur in the 12 direction perpendicular to the compression direction. At the passing on from stage II to stage III, the compression should be made along the O5 direction and the ex pansion along the perpendicular 23 direction. The next passage is accompanied by similar deformations of the plane in the O8 direction (compression) and in the perpendicular 35 direction (expansion). But we know from the prior consideration that the compression of a plane to any straight line with the coefficient k and the simultaneous expansion of a plane in the perpendicular direction with the same coefficient k are nothing but Hy perbolic Rotation. A scheme of hyperbolic transformation of the lattice fragment is presented in Fig. 5.19. The scheme corresponds to stage II of Fig. 5.18. Note that the hyperbola of the first quadrant has the equation xy=1 and the hyperbo la of the fourth quadrant has the equation xy=1. The transformation of the phyllotaxis lattice in the process of its growth is carried out by means of the hyperbolic rotation, the main yy xy = 1 geometric transformation of Y X hyperbolic geometry, follow ing from a consideration of the first key idea of Bodnar’s ge ometry. This transformation is accompanied by a modifica tion of dynamic symmetry, x which can be simulated by the sequential passage from the object with smaller sym metry order to the object xy = 1 with larger symmetry order. However, this idea does Figure 5.19. The general scheme of phyllotaxis lattice not give the answer to the transformation in a system of equatorial hyperboles question: why are the phyllo taxis lattices in Fig. 5.18 based on Fibonacci numbers? 5.12.4. The “Golden” Hyperbolic Functions For a more detailed study of the metric properties of the lattice in Fig. 5.19 we can examine its fragment represented in Fig. 5.20. Here the disposition of the points is similar to Fig. 5.19. Chapter 5 Hyperbolic Fibonacci and Lucas Functions 311 Let us pay attention to the basic peculiarities of the disposition of points in Fig. 5.20: (1) the points M1 and M2 are symmetrical relative to the bisector of the right angle YOX; (2) the geometric figures OM1M2N1, OM2N2N1, and OM2M3N2 are parallelograms; and (3) the point А is the vertex of the hyperbola xy=1, that is, xA=1, yA=1, therefore OA = 2. Y Let us evaluate the abscissa of point M2 denoted by x M2 = x . Tak ing into consideration the symmetry x of points M1 and M 2, we can write: x 1 x M1 = x −1 . It follows from the symme try condition of these points that the M √2 line segment M1M2 is tilted to the coor √2 M dinate axis under the angle of 45°. The M X O line segment M1M2 is parallel to the line 1 segment ON1; this means that the line N N segment ON1 is tilted to the coordinate axis under the angle of 45°. Therefore, the point N1 is the top of the lower branch of the hyperbola; here x N1 = 1, Figure 5.20. Analysis of the metric yN1 = 1, ON 1 = OA = 2 . It is clear properties of the phyllotaxis lattice that ON 1 = M1M 2 = 2 . It is obvious that the difference between the abscissas of the points M1 and M2 is equal to 1. These considerations result in the following equation for the calculation of the abscissa of the point M2, that is, x M2 = x : 1 1 2 3 2 1 x − x −1 = 1 or x 2 − x − 1 = 0. (5.155) This means that the abscissa x M2 = x is the positive root of the famous “gold en” algebraic equation (5.155): 1+ 5 . (5.156) 2 Thus, study of the metric properties of the phyllotaxis lattice in Fig. 5.20 un expectedly leads us to the golden mean. And this fact is the Second Key Outcome of Bodnar’s Geometry. This result was used by Bodnar for a detailed study of the phenomenon phyllotaxis. By developing this idea, Bodnar concluded that for a mathematical simulation of the phyllotaxis phenomenon we need to use a special class of the hyperbolic functions, the “Golden” Hyperbolic Functions: The “golden” hyperbolic sine x M2 = τ = Gshn = τn − τ− n . 2 (5.157) Alexey Stakhov THE MATHEMATICS OF HARMONY 312 The “golden” hyperbolic cosine τn + τ− n . (5.158) 2 Furthermore, Bodnar found a fundamental connection of the “golden” hy perbolic functions with Fibonacci numbers: 2 F2 k −1 = Gch ( 2k − 1) (5.159) 5 Gchn = 2 Gsh ( 2k ) . (5.160) 5 Using the correlations (5.159) and (5.160), Bodnar gave a very simple expla nation for the “puzzle of phyllotaxis”: why do Fibonacci numbers occur with such persistent constancy on the surface of phyllotaxis objects? The main reason is the fact that the geometry of “Living Nature,” in particular, the geometry of phyllo taxis, is a nonEuclidean geometry; however, this geometry differs substantially from Lobachevsky’s geometry and Minkowski’s fourdimensional world based on the classical hyperbolic functions. This difference consists in the fact that the main correlations of this geometry are described with the help of the “golden” hyperbol ic functions (5.157) and (5.158) that are connected with Fibonacci numbers by the simple correlations (5.159) and (5.160). It is important to emphasize that Bodnar’s model of the dynamic symme try of the phyllotaxis object illustrated in Fig. 5.20 is brilliantly confirmed by reallife phyllotaxis pictures of botanic objects. F2 k = 5.12.15. The Connection of Bodnar’s “Golden” Hyperbolic Functions with the Hyperbolic Fibonacci Functions Comparing the expressions for symmetric hyperbolic Fibonacci and Lucas sines and cosines given by formulas (5.57) (5.60) with expressions for Bodnar’s “golden” hyperbolic functions given by formulas (5.155) and (5.156), we discov er the following simple correlations between the indicated groups of the formu las: Gsh ( x ) = 5 sFs ( x ) 2 (5.161) Gch ( x ) = 5 cFs ( x ) 2 (5.162) Gsh ( x ) = 2sLs ( x ) (5.163) Chapter 5 Hyperbolic Fibonacci and Lucas Functions 313 Gch ( x ) = 2cLs ( x ) . (5.164) The analysis of these correlations allows us to conclude that the “golden” hy perbolic sine and cosine introduced by Oleg Bodnar [37], and the symmetric hyperbolic Fibonacci and Lucas sines and cosines introduced by Stakhov and Rozin [106], coincide within constant coefficients. The question of the use of the “golden” hyperbolic functions or the hyperbolic Fibonacci and Lucas func tions for the simulation of phyllotaxis objects has no particular significance because the final result is the same: it always results in an unexpected appear ance of Fibonacci or Lucas numbers on the surfaces of phyllotaxis objects. 5.13. Conclusion 1. The discovery of Lobachevsky’s geometry became an epochmaking event in the development not only mathematics, but also of science in general. Academician Kolmogorov appreciated the role of this discovery in the devel opment of mathematics in the following words [1]: “... It is difficult to overrate the importance of the reorganization of the entire warehouse of mathematical think ing, which happened in the 19th century. In this connection, Lobachevsky’s geometry was the most significant mathematical discovery at the start of the 19th century. Based upon this geometric insight, the belief in the absolute stabil ity of mathematical axioms was overthrown. This allowed for the creation of essen tially new and original abstract mathematical theories and, at last, to demonstrate that similar abstract theories can result hereafter in wider and more concrete ap plications.” After Lobachevsky’s discovery, the “hyperbolic ideas” started to pen etrate widely into various spheres of science. After the promulgation of the special theory of relativity by Einstein in 1905 and its “hyperbolic interpretation,” given by Minkowski in 1908, the “hyperbolic ideas” became universally recognized. Thus, a comprehension of the “hyperbolic character” of the processes in the physical world surrounding us became the major result in the development of science during the 19th and 20th centuries. The mathematical correlations of Lobachevsky’s geometry are or course based upon the classical hyperbolic functions. Why did Lobachevsky use these func tions, introduced by Vincenzo Riccati in the late 18th century, in his geometry? Apparently, Lobachevsky understood that these functions provide the best way to model the “hyperbolic character” of his geometry; however, he used them be Alexey Stakhov THE MATHEMATICS OF HARMONY 314 cause other hyperbolic functions at that moment simply did not exist. It is nec essary to note that Lobachevsky’s geometry, based on classical hyperbolic func tions, is historically the first “hyperbolic model” of physical space. Lobachev sky’s geometry and Minkowski’s geometry had put forward the hyperbolic func tions as the basic plan for modern science. 2. At the end of the 20th century, the Ukrainian architect Oleg Bodnar [37] and the Ukrainian mathematicians Alexey Stakhov and Ivan Tkachenko [98] broke the monopoly of classical hyperbolic functions in modern science. They introduced a new class of hyperbolic functions based on the golden mean. Later Alexey Stakhov and Boris Rozin developed the symmetrical hyperbolic Fibonacci and Lucas functions [108]. However, Bodnar, Stakhov, Tkachenko and Rozin each used their own unique methods to arrive at a new class of hyperbolic func tions. Oleg Bodnar found the “golden” hyperbolic functions thanks to his scien tific intuition, which resulted in the “golden” hyperbolic functions as the “key” idea in the study of the phyllotaxis phenomenon. These functions were used by him for the creation of the original geometric theory of phyllotaxis (Bodnar’s Geometry). The approach of Alexey Stakhov, Ivan Tkachenko and Boris Rozin was based on an analogy between the Binet formulas and hyperbolic functions. This approach resulted in the discovery of a new class of hyperbolic functions, Hyperbolic Fibonacci and Lucas Functions. 3. The hyperbolic Fibonacci and Lucas functions are a generalization of the Fibonacci and Lucas numbers “extended” to the continuous domain. There is a direct analogy between the Fibonacci and Lucas number theory and the theo ry of hyperbolic Fibonacci and Lucas functions because the “extended” Fibonac ci and Lucas numbers coincide with the hyperbolic Fibonacci and Lucas func tions at discrete values of the variable x (x=0, ±1, ±2, ±3, ...). Besides, every identity for the Fibonacci and Lucas numbers has its continuous analog in the form of a corresponding identity for the hyperbolic Fibonacci and Lucas func tions, and conversely. This outcome is of special significance for the Fibonacci number theory [13, 16, 28, 38] because this theory is as if it were transformed into the theory of hyperbolic Fibonacci and Lucas functions [98, 108, 116, 119]. Thanks to this approach, the Fibonacci and Lucas numbers became one of the most important numerical sequences of hyperbolic geometry. 4. Bodnar’s geometry [37, 52] demonstrates that the “golden” hyperbolic world exists objectively and independently of our consciousness and persis tently appears in Living Nature, in particular, in pine cones, pineapples, cacti, heads of sunflowers and baskets of various flowers in the form of Fibonacci and Lucas spirals on the surface of these biological objects (the phyllotaxis law). Furthermore Bodnar’s geometry is a demonstration of the “physical practica Chapter 5 Hyperbolic Fibonacci and Lucas Functions 315 bility” of the hyperbolic Fibonacci and Lucas functions, which underlie all of Living Nature. However, the promulgation of the new geometrical theory of phyllotaxis, made by the Ukrainian architect Oleg Bodnar [37], had demon strated that in addition to “Lobachevsky’s geometry.” Nature also uses other variants of the socalled “hyperbolic models of Nature” [98, 108]. The use of hyperbolic Fibonacci and Lucas functions allowed for the solution to the “rid dle of phyllotaxis,” that is, an explanation of, how Fibonacci and Lucas spirals appear on the surface of phyllotaxis objects. 5. However, perhaps the final step in the development of the “hyperbolic models” of Nature was made by Alexey Stakhov in his article [118]. The hyper bolic Fibonacci and Lucas mfunctions are a wide generalization of the symmet ric hyperbolic Fibonacci and Lucas functions introduced by Stakhov and Rozin in an earlier article [106]. They are based on the Gazale formulas and extend ad infinitum a number of new hyperbolic models of Nature. It is difficult to imagine that the set of new hyperbolic functions is infinite! The hyperbolic Fibonacci and Lucas mfunctions do complete a general theory of hyperbolic functions, started by Johann Heinrich Lambert and Vincenzo Riccati, and opens new per spectives for the development of new “hyperbolic ideas” in modern science. The development of a general theory of hyperbolic functions [118], based on the Gazale formulas [45], gives us the opportunity to put forward the fol lowing unusual hypothesis. Apparently, we can assume that theoretically there are an infinite number of “hyperbolic models of Nature.” One of the possible types of the hyperbolic functions given by the general formulas (5.105) (5.108) underlies these models. Each type of hyperbolic function meets some positive real number m, called the order of the function. This number generates new mathematical constants, the generalized golden mproportions, given by the formula (4.266). Each type of hyperbolic function, in turn, generates new classes of the recursive numerical sequences, the generalized Fibonacci and Lucas m numbers given by the recursive relations (4.251) and (4.302). Nature intelligently uses one or another type of hyperbolic function for introducing objects into this or that “hyperbolic world.” It is necessary to note that “Lobachevsky’s geometry” is one of the possible variants of the realization of the “hyperbolic world,” which probably is preferable for the objects of “min eral Nature.” Apparently in the case of “Living Nature,” “Bodnar’s geometry” based on the hyperbolic Fibonacci and Lucas functions, is preferable. This fact is confirmed by the “law of phyllotaxis” met in many botanical objects. But “Lobachevsky’s geometry” and “Bodnar’s geometry” are not unique variants of the realization of the “hyperbolic worlds.” It is possible to expect that new types of hyperbolic functions, described in [118], will result in the discovery of Alexey Stakhov THE MATHEMATICS OF HARMONY 316 new “hyperbolic worlds” which can be embodied in natural structures at dif ferent levels of organization in the Universe. Importantly, the occurrence of characteristic recursive numerical sequences given by the general recursive for mulas (4.251) and (4.302) are “external attributes”, embodied in Nature’s “hy perbolic worlds,” found in the search for corresponding types of hyperbolic functions (5.105) (5.108). Chapter 6 Fibonacci and Golden Matrices 317 Chapter 6 Fibonacci and Golden Matrices 6.1. Introduction to Matrix Theory 6.1.2. A History of Matrices The history of matrices goes back to ancient times. But the term “matrix” was not applied to the subject until 1850 by James Joseph Sylvester. “Matrix” is the Latin word for womb, and it retains that sense in English. It can also mean, more generally, any place in which something is formed or produced. The origin of mathematical matrices arose with the study of systems of simultaneous linear equations. An important Chinese text from between 300 BC and 200 AD, Nine Chapters of the Mathematical Art (Chiu Chang Suan Shu), is the first known example of the use of matrix methods to solve simultaneous equations. In the treatise’s seventh chapter, Too Much and Not Enough, the concept of a determinant first appears, nearly two millen nia before its supposed invention by the Japanese mathematician Seki Kowa in 1683, or alternatively, his German contemporary Gottfried Leibnitz (who is also credited with the invention of differential calculus, independently though simultaneously with Isaac Newton). More uses of matrixlike ar rangements of numbers appear in chapter eight, Methods of Rectangular Arrays, in which a method is given for solving simultaneous equations. This method is mathematically identical to the modern matrix method of solu tion outlined by Carl Friedrich Gauss (17771855), also known as Gauss ian elimination. James Joseph Sylvester (1814 – 1897) was an English mathematician. He made fundamental contributions to matrix theory, invariant theory, num ber theory, partition theory and combinatorics. He played a leading role in American mathematics in the latter half of the 19th century as a professor at Johns Hopkins University and as founder of the American Journal of Mathe matics. At his death, he was a professor at Oxford. Alexey Stakhov THE MATHEMATICS OF HARMONY 318 Sylvester began his study of mathematics at St. John’s College, Cambridge in 1831. In 1838 he be came professor of natural philosophy at University College London. In 1841 he moved to the United States to become a professor at the University of Virginia, but soon returned to England. On his re turn to England he studied law, alongside British lawyer/mathematician Arthur Cayley, with whom he made significant contributions to matrix theory. James Joseph Sylvester Poetry was one of Sylvester’s lifelong passions; he (1814 1897) read and translated works out of the original French, German, Italian, Latin and Greek, and many of his mathematical papers contain illustrative quotes from classical poetry. In 1870, following his ear ly retirement, Sylvester published a book titled The Laws of Verse in which he attempted to codify a set of poetry laws. In 1877 he again crossed the Atlantic to become the inaugural professor of mathematics at the new Johns Hopkins University in Baltimore, Maryland. In 1878 he founded the Amer ican Journal of Mathematics. And in 1883, Sylvester returned to England to become Savilian Professor of Geometry at Oxford University. Since their first appearance in ancient China, matrices have remained important mathematical tools. Today, they are used not only for solving systems of simultaneous linear equations, but also for describing quantum atomic structure, designing computer game graphics, analyzing relation ships, and in numerous other capacities. The elevation of the matrix from mere tool to important mathematical theory owes a lot to the work of female mathematician Olga Taussky Todd (19061995), who began by using matrices to analyze vibrations on airplanes and became the torchbearer for matrix theory. 6.1.2. Definition of a Matrix In mathematics, a Matrix (plural Matrices) is a rectangular table of numbers or, more generally, a table consisting of abstract quantities that can be added and multiplied. Matrices are used to describe linear equa tions, keep track of the coefficients of linear transformations, and to record data that depend on two parameters. Matrices can be added, multiplied, and decomposed in various ways, making them a key concept in linear algebra and matrix theory [157]. Chapter 6 Fibonacci and Golden Matrices 319 Let us consider, for example, a matrix A a11 a12 " a1n a a " a2n . A = 21 22 # # # # a a " a mn m1 m 2 The horizontal lines in a matrix are called Rows and the vertical lines are called Columns. A matrix with m rows and n columns is called an mbyn ma trix (written m×n) and m and n are called its Dimensions. The dimensions of a matrix are always given with the number of rows first, then the number of columns. The entry of a matrix A that lies in the ith row and the jth column is called the (i, j)th entry of A. This is written as Ai,j or A[i, j]. As is indicated, the row is always noted first, then the column. A matrix, where one of the dimensions is equal to 1, is often called a Vector. The (1×n)matrix (one row and n columns) is called a Row Vector, and the (m×1)matrix (m rows and one column) is called a Column Vector. 6.1.3. Matrix Addition and Scalar Multiplication Let A and B be two matrices with the same size, i.e. the same number of rows and of columns. The sum of A and B, written A+B is the matrix obtained by adding corresponding elements from A and B: a11 a12 ... a1n b11 b12 ... b1n a11 + b11 a12 + b12 ... a1n + b1n a21 a22 ... a2 n b21 b22 ... b2 n a21 + b21 a22 + b22 ... a2n + b2n ... ... ... ... + ... ... ... ... = ... . ... ... ... a a ... a b b ... b a + b a + b ... a + b m1 m 2 m1 m 2 m1 m1 m 2 m 2 mn mn mn mn The product of a scalar k and a matrix A, written kA or Ak is the matrix obtained by multiplying each element of A by k: a11 a12 ... a1n ka11 ka12 ... ka1n a21 a22 ... a2n ka21 ka22 ... ka2n k = . ... ... ... ... ... ... ... ... a a ... a ka ka ... ka mn mn m1 m2 m1 m 2 We also define: A=(1)A and AB=A+(B) Note that the matrix (A) is negative to the matrix A. Let us consider the following theorem for matrix addition and scalar multiplication. Alexey Stakhov THE MATHEMATICS OF HARMONY 320 Theorem 6.1. Let A, B and C be matrices with the same size and let k and k′ be scalars. Then we have: ( A + B ) + C = A + ( B + C ) , i.e., matrix addition is associative; ( A + B ) = (B + A), i.e., matrix addition is commutative; A + 0 = 0 + A; A + ( − A ) = 0; k ( A + B ) = kA + kB; ( k + k′) A = kA + k′ A; ( k × k′) A = k ( k′ A ) ; 1× A = A. 6.1.4. Matrix Multiplication Suppose A and B are two matrices such that the number of columns of A is equal to the number of rows of B, say A is an (m×p)matrix and B is a (p×n) matrix. Then the product of A and B, written AB is the (m×n)matrix whose (i, j)entry is obtained by multiplying the elements of the ith row of A by the corresponding elements of the jth column of B and then adding: a11 ... a1 p b11 ... b1 j ... b1n c11 ... c1n ... ... ... ... ... ... ... ... ... ... ... ai1 ... aip ... ... ... ... ... = ... cij ... , ... ... ... ... ... ... ... ... ... ... ... am1 ... amp bp1 ... bpj ... bpn cm1 ... cmn where p cij = ai1b1 j + ai 2b2 j + ... + aip bpj = ∑a b . ik kj k =1 Matrix multiplication satisfies the following properties: Theorem 6.2. ( AB)C = A ( BC ) ; A ( B + C ) = AB + AC ; k ( AB ) = ( kA) B = A ( kB ) , where k is a scalar. 6.1.5. Square Matrices 6.1.5.1. Definition of a Square Matrix A matrix with the same number of rows and columns is called a Square Matrix. A square matrix with n rows and n columns is said to be of order n, and is called an nsquare matrix. The Main Diagonal, or simply Diagonal, of a square matrix A=(aij) consists of the numbers: (a11, a22,…, ann) . Chapter 6 Fibonacci and Golden Matrices 321 The nsquare matrix with 1’s along the main diagonal and 0’s elsewhere, e.g. 1 0 0 0 I = 0 1 0 0 0010 0 0 0 1 is called the Unit or Identity Matrix and will be denoted by I. The unit matrix I plays the same role in matrix multiplication as the number 1 does in the usual multiplication of numbers. Specifically, AI=IA=A for any square matrix A. Square matrices can be raised to various powers. We can define powers of the square matrix A as follows: A2=AA, A3= A2A, …, and A0=I. 6.1.5.2. Invertible Matrices A square matrix A is said to be Invertible if there exists a matrix B with the property that AB=BA=I. Such a matrix B is unique; it is called the Inverse of A and is denoted by A1. 6.1.5.3. Determinants To each nsquare matrix A=(aij) we assign a specific number called the Determinant of A, denoted by Det(A) or A. The determinants of order one, two and three are defined as follows: Det ( a11 ) = a11 = a11 a a Det 11 12 = a11a22 − a12a21 a 21 a22 a11 a12 a13 a21 a22 a23 = a11a22 a33 + a12 a23 a31 + a13 a21a32 − a13 a22 a31 − a12 a21a33 − a11a23 a32 a31 a32 a33 An important determinant property is given by the following theorems. Theorem 6.3. For any two nsquare matrices A and B we have Det(A×B)=DetA×DetB Theorem 6.4 Det(An)=(DetA)n. (6.1) Alexey Stakhov THE MATHEMATICS OF HARMONY 322 6.2. Fibonacci QMatrix 6.2.1. A Definition of the QMatrix In recent decades the theory of Fibonacci numbers [13, 16, 28] has been sup plemented by the theory of the socalled Fibonacci Qmatrix [9]. The latter is the (2×2)matrix of the following form: ( ) Q= 11 . 10 (6.2) Note that the determinant of the Qmatrix is equal to 1: DetQ=1. (6.3) The article [158] devoted to the memory of Verner E. Hoggatt, founder of the Fibonacci Association, contained the history and an extensive bibliogra phy of the Qmatrix and emphasized Hoggatt’s contribution in its develop ment. Although the name of the Qmatrix was introduced before Verner E. Hoggatt, it was from Hoggatt’s articles that the idea of the Qmatrix “caught on like wildfire among Fibonacci enthusiasts. Numerous papers have appeared in ‘The Fibonacci Quarterly’ authored by Hoggatt and/or his students and other collaborators where the Qmatrix method became a central tool in the analysis of Fibonacci properties” [158]. We will consider here a theory of the Qmatrix developed in Hoggatt’s book [16]. 6.2.2. Properties of the QMatrix The following theorem connects the Qmatrix to Fibonacci numbers. Theorem 6.5. For a given integer n the nth power of the Qmatrix is given by F F Q n = n +1 n , F F n n −1 where Fn1, Fn, Fn+1, are Fibonacci numbers. Proof. We use mathematical induction. Clearly, for n=1, F F Q 1 = 2 1 = 1 1 . F1 F0 1 0 F F k Suppose that Q = Fk +1 F k , then k k −1 (6.4) Chapter 6 Fibonacci and Golden Matrices 323 F 1 1 F + F F F F F Q k +1 = Q k ×Q = k +1 k × = k +1 k k +1 = k + 2 k +1 . Fk Fk −1 1 0 Fk + Fk −1 Fk Fk +1 Fk The theorem is proved. The next theorem gives a formula for the determinant of the matrix (6.5). Theorem 6.6. For a given integer n we have: (6.5) Det(Qn)=(1)n. Proof. Using (6.1) and (6.5), we can write: Det(Qn)=( DetQ)n=(−1)n. The theorem is proved. The following remarkable property for Fibonacci numbers follows from Theorem 6.5: n (6.6) DetQ n = Fn −1Fn +1 − Fn2 = ( −1) . Recall that the identity (6.6) is one of the most important identities for Fibonacci numbers. This one is called the Cassini formula in honor of the famous French astronomer Giovanni Domenico Cassini (16251712), who discovered this formula for the first time. Now, represent the matrix (6.4) in the following recursive form: F + Fn −1 Fn −1 + Fn −2 Fn Fn −1 Fn −1 Fn − 2 (6.7) Qn = n = + Fn −1 + Fn −2 Fn −2 + Fn −3 Fn −1 Fn −2 Fn − 2 Fn − 3 or (6.8) Q n = Q n −1 + Q n −2 . We can represent the recursive relation (6.8) in the following form: n −2 (6.9) Q = Q n − Q n −1. n n The explicit forms of the matrices Q and Q ( n = 0, ±1, ±2, ±3,... ) obtained by means of the use of the recursive relations (6.8) and (6.9) are given in Table 6.1. Table 6.1. Qmatrices n Q n Q −n 0 1 1 1 1 0 0 1 1 −1 1 2 3 4 5 ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) 1 0 1 0 0 1 0 1 2 1 1 −1 1 1 −1 2 3 2 −1 2 2 1 2 −3 5 3 2 −3 3 2 −3 5 8 5 −3 5 5 3 5 −8 6.2.3. Binet Formulas for the QMatrix We can find the analytical expressions for the following sum and difference: Qn+Qn and Qn−Qn. (6.10) Alexey Stakhov THE MATHEMATICS OF HARMONY 324 We can start from the following examples for the even (n=2k) and odd (n=2k+1) powers. For the case n=5 we have the following expression for the matrix sum (6.10): ( )( )( ) ( ) Q 5 + Q −5 = 8 5 + −3 5 = 5 10 = 5 1 2 = F5T , 5 3 5 −8 10 −5 2 −1 (6.11) where F5 is a Fibonacci number and ( ) T= 1 2 . 2 −1 (6.12) We have the following expression for the matrix difference (6.10): ( )( )( ) ( ) Q 5 − Q −5 = 8 5 − −3 5 = 11 0 = 11 1 0 = L5 I , 5 3 5 −8 0 11 0 1 (6.13) where L5 is a Lucas number, I is the identity matrix. For the case n=6 we have: ( )( )( ) ( ) Q − Q = (13 8 ) − ( 5 −8 ) = ( 8 16 ) = 8 ( 1 2 ) = F T , −8 13 8 5 16 −8 2 −1 Q 6 + Q −6 = 13 8 + 5 −8 = 18 0 = 18 1 0 = L6 I −8 13 8 5 0 18 0 1 6 −6 6 (6.14) (6.15) where the matrix T is given by (6.12). The expressions (6.11), (6.13), (6.14), (6.15) are the partial cases of the following general formulas valid for an arbitrary n (n=0, ±1, ±2, ±3,…): Q2k+1 Q(2k+1)=L2k+1I (6.16) Q2k Q2k=L2kI (6.17) Q2k+1+ Q(2k+1)=F2k+1T (6.18) Q2k+ Q2k=F2k+1T (6.19) Note that the formulas (6.16) (6.19) are the matrix equivalents of the Binet formulas (2.67) and (2.68). The formulas (6.16) and (6.17) are the ma trix equivalent of the Binet formula (2.67) for Lucas numbers and the formu las (6.18) and (6.19) are the matrix equivalents of the Binet formula (2.68) for Fibonacci numbers. It is clear that the Qmatrix (6.2) in these formulas plays the role of the golden ratio τ in the formulas (2.67) and (2.68). For all that, the special matrix (6.12) plays the role of the irrational number 5 in the Binet formula (2.68) for Fibonacci numbers. This analogy between the golden ratio τ and Qmatrix (6.2) shows that there is an isomorphism between the golden mean theory and the Q matrix theory. We can prove the following theorems confirming this isomor Chapter 6 Fibonacci and Golden Matrices 325 phism. First of all, we can write the following trivial property for the golden ratio powers: τnτm =τmτn =τn+m . Note that we do not need to prove for the number τ the following obvious equality: τnτm =τmτn, however, for the matrices we need to prove the similar equality. We can prove the following theorem for the Qmatrices. Theorem 6.7. QnQm =QmQn =Qn+m . Proof. At first, we prove the identity (6.20) QnQm = Qn+m . In order to prove this identity we can write the product of the matrices QnQm as follows: Fn Fm +1 Fm Fn +1 Q nQ m = × F Fm Fm −1 F n −1 n F F +F F = n +1 m +1 n m Fn Fm +1 + Fn −1Fm Fn +1Fm + Fn Fm −1 . Fn Fm −1 + Fn −1Fm −1 (6.21) For the proof we use the formula for the generalized Fibonacci numbers Gn described in Section 2.5: Gn+m=Gn+1Fm+GnFm1 . If Gi=Fm , the formula (6.22) takes the following form: (6.22) (6.23) Fn+m=Fn+1Fm+FnFm1. Using (6.23), we can represent the elements of the matrix (6.21) as follows: Fn+1Fm+1+FnFm= Fn+m+1 (6.24) Fn+1Fm+FnFm1= Fn+m (6.25) FnFm+1+Fn1Fm= Fn+m (6.26) FnFm1+Fn1Fm1= Fn+m1. (6.27) Taking into consideration the formulas (6.24) (6.27), we can represent the matrix (6.21) as follows: F F Q nQ m = n + m +1 n + m = Q n + m . Fn + m Fn + m −1 By analogy we can prove the validity of the following identity: QmQn = Qn+m . Alexey Stakhov THE MATHEMATICS OF HARMONY 326 It follows from the identity (6.20) that the matrices Qn and Qm possess the property of multiplication commutativity. Thus, the above property of the Qmatrix, given by (6.20), gives us the right 1 1 to affirm that the Fibonacci matrix Q = 1 0 is a special matrix, which plays in the theory of twobytwo matrices a particular role similar to the role of the golden mean in the theory of real numbers. ( ) 6.3. Generalized Fibonacci QpMatrices 6.3.1. A Definition of the QpMatrices Note that the Qmatrix is a generating matrix for the classical Fibonacci numbers given by the recursive relation (2.3). In the article [103] Alexey Stakhov generalized the concept of the Qmatrix [16, 158] and introduced the generalized Fibonacci Qpmatrices. For a given p (p=1, 2, 3, …) the gener alized Fibonacci Qpmatrix has the following form: 1 1 0 0 " 0 0 0 0 1 0 " 0 0 0 0 0 1 " 0 0 Qp = # # # # # # # 0 0 0 0 " 1 0 (6.28) 0 0 0 0 " 0 1 1 0 0 0 " 0 0 The Qpmatrix (6.28) is a square (p+1)×(p+1)matrix. It consists of the identity (p×p)matrix bordered by the last row consisting of 0’s and the lead ing 1, and the first column consisting of 0’s embraced by a pair of 1’s. We can list the Qp matrices for the case p=1, 2, 3, … as follows: 1 1 0 0 0 1 1 0 0 1 1 0 0 0 1 0 0 11 0 0 1 0 Q1 = ; Q4 = 0 0 0 1 0 . = Q ; Q2 = 0 0 1 ; Q 3 = 10 0 0 0 1 1 0 0 0 0 0 0 1 1 0 0 0 1 0 0 0 0 ( ) (6.29) 6.3.2. The Main Theorems for the QpMatrices Next let us raise the Qpmatrix (6.28) to the nth power and find the analytical expression for the matrix Q pn . Let us prove the following theorem. Theorem 6.8. For a given integer p= 1, 2, 3, … we have: Chapter 6 Fibonacci and Golden Matrices 327 Fp (n) " Fp (n − p + 2) Fp (n − p + 1) Fp (n + 1) Fp (n − p + 1) Fp (n − p) " Fp (n − 2 p + 2) Fp (n − 2 p + 1) , Q pn = # # # # # F (n − 1) F (n − 2) " F (n − p) (6.30) Fp (n − p − 1) p p p Fp (n − 1) " Fp (n − p + 1) F p (n − p ) F p (n) where Fp(n) is the Fibonacci pnumber, n=0, ±1, ±2, ±3, … . Before the proof of Theorem 6.8 we analyze the matrix (6.30). Note that all entries of the matrix are Fibonacci pnumbers given by the recursive relation (4.18) at the seeds (4.19). The matrix (6.30) consists of (p+1) rows and (p+1) columns. The first row is the following sequence of the Fibonacci pnumbers: Fp(n+1), Fp(n),…, Fp(np+2), Fp(np+1), (6.31) the second row consists of the Fibonacci pnumbers: (6.32) Fp(np+1), Fp(np),…, Fp(n2p+2), Fp(n2p+1), the pth row consists of the Fibonacci pnumbers: (6.33) Fp(n1), Fp(n2),…, Fp(np), Fp(np1), the (p+1)th row consists of the following Fibonacci pnumbers: (6.34) Fp(n), Fp(n1),…, Fp(np+1), Fp(np). Note that every one of the rows given by (6.31) – (6.34) is the sequence of the Fibonacci pnumbers consisting of (p+1) sequential Fibonacci pnumbers. The first row (6.31) begins with the Fibonacci pnumber Fp(n+1) and finishes with the Fibonacci pnumber Fp(np+1), the second row (6.32) begins with the Fibonacci pnumber Fp(np+1) and finishes with the Fibonacci pnumber Fp(n 2p+1) and finally, the (p+1)th row begins from the Fibonacci pnumber Fp(n) and finishes with the Fibonacci pnumber Fp(np). It is important to emphasize that the second Fibonacci pnumber of the first row (6.31) and the first Fi bonacci pnumber of the last row (6.34) are equal to Fp(n), the third Fibonacci pnumber of the second row (6.32), the fourth Fibonacci pnumber of the third row, ..., and the (p+1)th Fibonacci pnumber of the pth row (6.33) are equal to Fp(np1). Now we can start a proof of Theorem 6.8. Proof. For a given p we can use the induction method. (a) The basis of the induction. At first we prove the basis of the induction, that is, we prove that the statement (6.30) is valid for the case n=1. It is clear that for the case n=1 the matrix (6.30) takes the following form: Fp (1) " Fp (− p + 3) Fp (− p + 2) Fp (2) Fp (− p + 2) Fp (− p + 1) " Fp (−2 p + 3) Fp (−2 p + 2) . # # # # # F (0) ( − 1 ) " ( − + 1 ) ( − ) F F p F p p p p p Fp (0) " Fp (− p + 2) Fp (− p + 1) Fp (1) (6.35) Alexey Stakhov THE MATHEMATICS OF HARMONY 328 Taking into consideration the values of the initial terms of the Fibonacci p numbers given by (4.19) and the mathematical properties of the extended Fi bonacci pnumbers given by (4.29), (4.31) and (4.32) we can write: Fp(2)=Fp(1)=Fp(p)=1. Note that all the remaining entries of the matrix (6.34) are equal to 0. This means that the matrix (6.35) coincides with the Qpmatrix given by (6.28). The basis of the induction is proved. (b) Inductive hypothesis. Suppose that the statement (6.30) is valid for arbitrary n, and then prove it for n=1. For this purpose we shall consider the product of the matrices 1 1 0 0 " 0 0 Fp ( n) " Fp ( n − p + 2) Fp ( n − p + 1) Fp (n + 1) 0 0 1 0 " 0 0 Fp ( n − p + 1) Fp ( n − p) " Fp (n − 2 p + 2) Fp (n − 2 p + 1) 0 0 0 1 " 0 0 n Qp ×Qp = # # # # # × # # # # % # # . (6.36) Fp (n − 1) Fp (n − 2) " Fp (n − p) Fp (n − p − 1) 0 0 0 0 " 1 0 F (n) Fp ( n − 1) " Fp ( n − p + 1) Fp ( n − p ) 0 0 0 0 " 0 1 p 1 0 0 0 " 0 0 Let us now consider the matrix (6.36). If we multiply the first row of the matrix Q pn by the matrix Qp in the example (6.36) we obtain the entry a11: a11= Fp(n+1)+ Fp(np+1)= Fp(n+2). Continuing the process of the matrix multiplication for the example (6.36), we obtain the following entries of the first row of the matrix Q pn+1 : Fp(n+2), Fp(n+1),…, Fp(np+3), Fp(np+2). (6.37) Next let us compare the first row of the matrix Q pn given by (6.31) with the first row of the matrix Q pn+1 given by (6.37). We can see that all arguments of the Fibonacci pnumbers in the sequence (6.37) differ by 1 from the arguments of the corresponding Fibonacci pnumbers, which form the first row (6.31) of the matrix (6.30). By analogy we can show that this rule is valid for all entries of the matrix Q pn+1 , which is formed from the matrix Q pn by means of its multipli cation by the matrix Qp in accordance with (6.36). This consideration proves the validity of Theorem 6.8. It is easy to prove the following identity for the product of the following Qp matrices: Q pn +1 = Q pn × Q p = Q p × Q pn . Using (6.38), we can prove the following theorem. Theorem 6.9. Q pn × Q pm = Q pm × Q pn = Q pm + n . (6.38) Chapter 6 Fibonacci and Golden Matrices 329 Decomposing each entry of the matrix (6.30) the Fibonacci pnumber in accordance with the basic recursive relation (4.18), we obtain the following re sult given by Theorem 6.10. Theorem 6.10. Q pn = Q pn −1 + Q pn − p −1. (6.39) This recursive relation (6.39) can also be represented in the following form: Q pn − p −1 = Q pn − Q pn −1 . (6.40) n For example, consider the matrices Q p corresponding to the cases p=2 and p=3 F2 (n + 1) n Q2 = F2 (n − 1) F2 (n) F2 (n − 1) F2 (n − 2) F2 (n − 3) F2 (n − 1) F2 (n − 2) F2 (n) F3 (n + 1) F3 (n) F3 (n − 1) F3 (n − 2) F (n − 2) F (n − 3) F (n − 4) F (n − 5) 3 3 3 Q3n = 3 . F n F n F n F ( − 1 ) ( − 2 ) ( − 3 ) (n − 4) 3 3 3 3 F3 (n) F3 (n − 1) F3 (n − 2) F3 (n − 3) (6.41) (6.42) Using the matrices (6.41) and (6.42) and the general matrix (6.30), we obtain some special matrices, in particular, the matrices of the kind Q pp and the inverse matrices of the kind Q p−1. For example, for the cases p=2, 3 , 4, 5 the matrices of the kind Q pp and Q p−1 have the following form: p=2 F2 (3) 2 Q2 = F2 (1) F2 (2) F2 (1) 1 1 1 F2 (0) F2 (−1) = 1 0 0 F2 (2) F2 (1) F2 (0) 1 1 0 F2 (0) F2 (−1) F2 (−2) 0 0 1 Q2−1 = F2 (−2) F2 (−3) F2 (−4) = 1 0 −1 F2 (−1) F2 (−2) F2 (−3) 0 1 0 p=3 F3 (4) F3 (3) F3 (2) F3 (1) 1 1 1 1 F (1) F (0) F (−1) F (−2) 1 0 0 0 3 3 3 Q33 = 3 = F F F ( 2 ) ( 1 ) ( 0 ) 1 1 0 0 F 3 3 3 3 (−1) F3 (3) F3 (2) F3 (1) F3 (0) 1 1 1 0 F3 (0) F3 (−1) F3 (−2) F3 (−3) 0 0 0 1 F (−3) F (−4) F (−5) F (−6) 1 0 0 −1 3 3 3 Q3−1 = 3 = 2 ( ) F F F F − ( − 3 ) ( − 4 ) 0 1 0 0 3 3 3 (−5) 3 F3 (−1) F3 (−2) F3 (−3) F3 (−4) 0 0 1 0 (6.43) (6.44) (6.45) (6.46) Alexey Stakhov THE MATHEMATICS OF HARMONY 330 p=4 0 0 0 0 1 1 1 1 1 1 1 0 0 0 0 1 0 0 0 −1 Q44 = 1 1 0 0 0 Q4−1 = 0 1 0 0 0 1 1 1 0 0 0 0 1 0 0 1 1 1 1 0 0 0 0 1 0 (6.47) p=5 1 1 1 1 1 1 0 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0 0 −1 1 1 0 0 0 0 −1 0 1 0 0 0 0 5 . Q5 = Q = 1 1 1 0 0 0 5 0 0 1 0 0 0 (6.48) 1 1 1 1 0 0 0 0 0 1 0 0 1 1 1 1 1 0 0 0 0 0 1 0 Looking at the matrices (6.43) (6.48), we can see that they have a strong regular form. This allows us to construct similar matrices for arbitrary p. Comparing the matrices of the kind Q pp ( p = 1, 2, 3, ... ) we can see that each ma trix Q pp includes in itself all the preceding matrices of the kind Q pp−−11 , Q pp−−22 , ..., Q11. For example, the matrix Q pp−−11 is built up from the matrix Q pp by means of strik ing out the last row and last column of the matrix Q pp . The last rows and columns of all matrices of the kind Q pp have a strong regular structure. In particular, each last row has the following form (111…10) (p+1 entries) and each last column begins with the top 1, but all the remain ing entries are equal to zero. Comparing the inverse matrices of the kind Q p−1 ( p = 1, 2, 3, ... ) , we can find the following regularity. The main diagonal of the matrix Q p−1 consists of zeros. All entries, forming the diagonal under the main diagonal, are equal to 1. The last column starts with the entries 1 and −1. All of the remaining entries of the ma trix Q p−1 are equal to 0. −1 There is the following connection between the next matrices Q p−1 and Q p− 1. −1 −1 The matrix Q p−1 can be obtained from the matrix Q p , if we strike out the last row and the next to the last column of the matrix Q p−1 . 6.4. Determinants of the QpMatrices and their Powers Now let us calculate the determinants of the Qpmatrices (6.28). It is clear that DetQ1=1. Calculate the determinant of the Q2matrix 1 1 0 Q2 = 0 0 1 . 1 0 0 (6.49) Chapter 6 Fibonacci and Golden Matrices 331 Comparing the Q2matrix (6.49) with the Q1matrix Q1 = 1 1 = Q , (6.50) 1 0 we can see that the Q2matrix (6.49) is reduced to the Q1matrix (6.50) if we strike out from the Q2matrix (6.49) the 2nd row and the 3rd column. Note that the sum of these numbers 2+3=5. According to the matrix theory [157] we have: ( ) DetQ2=1×(1)5×DetQ1=1×(1)×(1)=1. Now let us calculate the determinant of the Q3matrix (6.51) 1 1 0 0 Q3 = 0 0 1 0 . 0 0 0 1 (6.52) 1 0 0 0 Comparing the matrix (6.52) with the matrix (6.49) we can see that the Q3 matrix (6.52) is reduced to the Q2 matrix (6.49) if we strike out the 3rd row and the 4th column (3+4=7) in the matrix (6.52). This means [157] that DetQ3=1×(1)7×DetQ2=1×(1)×1=1. (6.53) By analogy we can show that DetQ4=1×(1)9×DetQ3=1×(1)×(1)=1. (6.54) Ccontinuing this process, that is, generating the formulas similar to (6.51), (6.53), and (6.54) we can write the following recursive relation, which con nects the determinants of the adjacent matrices Qp and Qp1: DetQp=1×(1)2p+1×DetQp1. (6.55) It is easy to show that DetQp is equal to 1 for the even values of p=2k and is equal to (1) for the odd values of p=2k+1. We can formulate this result in the form of the following theorem. Theorem 6.11. For a given p=1,2,3,… we have: DetQp=(1)p . (6.56) Let us calculate the determinants of the inverse matrices of the kind Q p−1 . For this purpose we consider the wellknown correlation, which connects the “di rect” and “inverse” matrices: Q p−1 × Q 1p = I , (6.57) where I is the identity matrix with DetI=1. (6.58) Applying to (6.57) the wellknown rule given by Theorem 6.3 for the deter minant of the product of two matrices and taking into consideration (6.58), we can write: Alexey Stakhov THE MATHEMATICS OF HARMONY 332 Det Q p−1 × Det Q 1p = Det I = 1. (6.59) In accordance with (6.56) the determinant of the matrix Q 1p takes one of the two values, 1 or 1; it follows from (6.59) that Det Q p−1 = Det Q 1p = 1. (6.60) Let us calculate the determinant of the matrix Q pn using (6.1): n Det Q pn = Det Q p , (6.61) ( ) where n=0,±1,±2,±3,… . Taking into consideration the result of Theorem 6.11, we can formulate the following theorem. Theorem 6.12. For given p=1,2,3,… and n=0,±1,±2,±3,…, the determi nant of the matrix Q pn is given by the following expression: (6.62) Det Q pn = ( −1) . It is clear that Theorem 6.12 is a generalization of the wellknown Theorem 6.6 for the Qmatrix that corresponds to the case p=1. Note that Theorems 6.8 and 6.12 are a source for new results in the field of Fibonacci number theory [16]. For example, consider the matrix Q2n for the case p=2 pn F2 (n + 1) F2 (n) F2 (n − 1) Q2n = F2 (n − 1) F2 (n − 2) F2 (n − 3) F2 (n) F2 (n − 1) F2 (n − 2) (6.63) where n=0,±1,±2,±3,… . By calculating the determinant of the matrix (6.63) and by using the identity (6.62), we can write the case for p=2: Det Q2n = 1. (6.64) It is easy to prove the following theorem. Theorem 6.13. For p=2 and n=0,±1,±2,±3,…, we have the following iden tity connecting the adjacent Fibonacci 2numbers: Det Q2n = F2 ( n + 1) F2 ( n − 2) F2 ( n − 2 ) − F2 ( n − 1) F2 ( n − 3 ) + F2 ( n ) F2 ( n ) F2 ( n − 3 ) − F2 ( n − 1) F2 ( n − 2 ) + F2 ( n − 1) F2 ( n − 1) F2 ( n − 1) − F2 ( n ) F2 ( n − 2 ) = 1. (6.65) It is clear that Theorem 6.8 and 6.12 give a theoretically infinite number of correlations similar to (6.6) and (6.65). If we remember that Fibonacci pnum bers were obtained in the study of the diagonal sums of Pascal’s triangle, we can assert that the identities similar to (6.6) and (6.65) express some unusual math ematical properties of Pascal’s triangle! Chapter 6 Fibonacci and Golden Matrices 333 6.5.The “Direct” and “Inverse” Fibonacci Matrices Again, let us consider Table 6.1, which gives the “direct” and “inverse” Q matrices. By comparing the “direct” (Qn) and “inverse (Qn) Fibonacci Q matrices, it is easy to find a very simple method that allows us to obtain the “inverse” matrix Qn from its “direct” matrix Qn. In fact, if the power n of the “direct” matrix Qn given by (6.4) is even (n=2k), then for obtaining its inverse matrix Qn it is simply necessary to interchange the places of the diagonal elements Fn+1 and Fn1 in (6.4) and to change the sign of the diagonal elements Fn. This means that for the case n=2k the “inverse” matrix Qn has the following form: − F2 k F Q −2 k = 2 k −1 . (6.66) − F2 k F2 k +1 In order to obtain the “inverse” matrix Qn from the “direct” matrix Qn given by (6.4) for the case n=2k+1, it is necessary to interchange the places of the diagonal elements Fn+1 and Fn1 in (6.4) and give them the opposite sign, that is: F2 k +1 −F Q −2 k −1 = 2 k . (6.67) F2 k +1 − F2 k +2 Another way to obtain the matrices Qn follows directly from the expression (6.4). Here we can represent two Fibonacci series Fn+1 and Fn1 shifted by one number with respect to the other (Table 6.2). If we select the num Table 6.2. Fibonacci series Fn+1 and Fn1 ber n=1 in the first row n 5 4 3 2 1 0 −1 −2 −3 −4 5 −6 of Table 6.2 and then se lect the four Fibonacci 1 −1 2 3 5 Fn +1 8 5 3 2 1 1 0 numbers in the lower Fn 5 3 2 1 1 0 1 −1 2 −3 5 8 two rows under the num ber 1 and to the right with respect to it, then the totality of these Fibonacci numbers form the Qmatrix. The Qmatrix is singled out in bold type in Table 6.2. If we move in Table 6.2 to the left with respect to the Qmatrix, then we obtain the matrices Q2, Q3,…, Qn, respectively. If we move in Table 6.2 to the right with respect to the Qmatrix, then we obtain the matrices Q0, Q1,…, Qn, respectively. Also the Fibonacci matrices Q5 and the “inverse” to it, Fibonac ci matrix Q5 are singled out in bold type in Table 6.2. Note that the matrix Q0 is an identity matrix. This principle of the Qmatrices design can be used for the general case of Qpmatrices. The analysis of the matrix (6.30) for the case p=2 shows that all THE MATHEMATICS OF HARMONY Alexey Stakhov 334 matrices of the kind Q2n can be obtained from (6.30), provided we represent three series of the adjacent Fibonacci 2numbers F2(n+1), F2(n1), F2(n), shift ed with respect to each other as is shown in Table 6.3. Table 6.3. Fibonacci 2numbers F2(n+1), F2(n1), F2(n) n 5 4 3 2 1 0 −1 −2 −3 −4 5 −6 −7 F2 ( n + 1) 4 3 2 1 1 1 0 0 1 0 1 1 1 2 1 1 1 0 0 1 0 −1 1 1 2 0 3 2 1 1 1 0 0 1 0 −1 1 1 2 F2 ( n − 1) F2 ( n ) Select the number n=1 in the first row of Table 6.3 and then select the known Fibonacci 2numbers in the lower rows with respect to the number 1 as is shown in Table 6.3 (they are singled out by bold type). It is clear that the totality of singled out Fibonacci 2numbers form the Q2matrix. If we move to the left in Table 6.3 with respect to the Q2matrix, then we obtain the matrices Q22 , Q23 , ..., Q2n , respectively. If we move to the right in Table 6.3 with respect to the Q 2matrix, then we obtain the matrices Q20 = I , Q2−1, ..., Q2− n , respectively. By analogy, using (6.30), we can construct the table of the Fibonacci 3numbers that give the matrices of the kind Q3n (see Table 6.4). Table 6.4. Fibonacci 3numbers F3(n+1), F3(n2),F3(n1), F3(n) n 5 4 3 2 1 0 −1 −2 −3 −4 −5 −6 −7 F3 ( n + 1) 3 2 1 1 1 1 0 0 0 1 0 0 −1 1 1 1 0 0 0 1 0 0 −1 1 0 1 1 1 1 1 0 0 0 1 0 0 −1 1 0 2 1 1 1 1 0 0 0 1 0 0 −1 1 F3 ( n − 2) F3 ( n − 1) F3 ( n ) 6.6. Fibonacci GmMatrices 6.6.1. A Definition of the Fibonacci GmMatrix In Chapter 4 we introduced the generalized Fibonacci mnumbers given by the recursive relation F m(n+2)=mFm(n+1)+Fm(n) at the seeds (6.68) Chapter 6 Fibonacci and Golden Matrices 335 Fm(0)=0, Fm(1)=1, (6.69) where m>0 is a given real number and n=0,±1,±2,±3,… . Similar to the Fibonacci Qmatrix (6.2), which is a generating matrix for the classical Fibonacci numbers, we can introduce the Gmmatrix [118] that is a generating matrix for the Fibonacci mnumbers given by the recursive relation (6.68) at the seeds (6.69). Gmmatrix ( ) Gm = m 1 . 1 0 (6.70) Note that the determinant of the Gmmatrix (6.70) is equal to −1: Det Gm =m×0 −1×1=−1. (6.71) The following theorem gives a connection of the Gmmatrix (6.70) to the Fibonacci mnumbers given by (6.68) and (6.69) . Theorem 6.14. For a given integer n=0,±1,±2,±3,…, the nth power of the Gmmatrix is given by F (n + 1) Fm (n) Gmn = m , Fm (n) Fm (n − 1) where Fm(n−1), Fm(n), Fm(n+1) are the Fibonacci mnumbers. Proof. We use mathematical induction. Clearly, for n=1 (6.72) F (2) Fm (1) Gm1 = m (6.73) . Fm (1) Fm (0) By using the seeds (6.69) and the recursive relation (6.68), we can write: Fm(0)=0, Fm(1)=1, Fm(2)=mFm(1)+ Fm(0)=m. It follows from (6.73) and (6.74) that ( ) Gm1 = m 1 . 1 0 (6.74) (6.75) The basis of the induction is therefore proved. Suppose that for a given integer k our inductive hypothesis is the following: F k +1 Fmk Gmk = m k . k −1 Fm Fm We can then write: Fm (k) m 1 F (k + 1) Gmk+1 = Gmk × Gm = m × Fm (k − 1) 1 0 Fm (k) Alexey Stakhov THE MATHEMATICS OF HARMONY 336 mF ( k + 1) + Fm ( k ) Fm ( k + 1) Fm ( k + 2 ) Fm ( k + 1) = m = . mF ( k ) + F ( k − 1) Fm ( k ) Fm ( k + 1) Fm ( k ) m m Thus, the theorem is proved. 6.6.2. Determinants of the GmMatrices The next theorem gives a formula for the determinant of the matrix (6.72). Theorem 6.15. For a given integer n=0,±1,±2,±3,…, we have: Det Gmn = (−1) . n (6.76) Proof. By using the property (6.1) and by taking into consideration (6.71), we can write: Det Gmn = ( Det Gm ) = (−1) . Thus, the theorem is proved. n n 6.6.3. GmMatrix and Cassini Formula for the Fibonacci mNumbers If we calculate the determinant of the matrix (6.72) and take into consid eration (6.76), we obtain the following identity: (6.77) Det Gmn = Fm (n + 1)× Fm (n −1)− Fm2 (n) = (−1) . Note that the identity (6.77) is one of the most important identities for the Fibonacci mnumbers. It is clear that the identity (6.77) is a generaliza tion of the wellknown Cassini formula (6.6). n 6.6.4. Some Properties of the GmMatrices Theorem 6.16. Gmn = mGmn−1 + Gmn−2 . (6.78) Proof. By using the recursive relation (6.68), we can represent the matrix (6.72) as follows: mFm (n − 1) + Fm (n − 2) mFm (n) + Fm (n − 1) Gmn = mFm (n − 1) + Fm (n − 2) mFm (n − 2) + Fm (n − 3) F (n) Fm ( n − 1) Fm ( n − 1) Fm ( n − 2 ) + = mGmn−1 + Gmn−2 . = m m F ( n − 1) F ( n − 2 ) F ( n − 2 ) F ( n − 3 ) m m m m Thus, the theorem is proved. Chapter 6 Fibonacci and Golden Matrices 337 We can also represent the recursive relation (6.78) in the following form: Gmn−2 = Gmn − Gmn−1 . (6.79) Based on the recursive relations (6.78) and (6.79), we can construct the se quences of the Gmmatrices (6.72) similar to Table 6.1. Note that for the case m=1 the matrices G1n coincide with the matrices Qn, that is, Table 6.1 gives a sequence of the matrices G1n . Consider the case m=2. Remember that for this case a sequence of the Fibonacci mnumbers Fm=2(n) looks similar to Table 6.5. Table 6.5. A sequence of the Fibonacci mnumbers for the case m=2 n 6 5 4 3 2 1 0 −1 −2 −3 −4 −5 −6 F2 ( n ) 70 29 12 5 2 1 0 1 −2 5 −12 29 −70 Let us construct a sequence of the matrices G2n . For the case n=0, we define the matrix Gmn as follows: ( ) Gm0 = 1 0 . 0 1 (6.80) Using the recursive relation (6.78) and taking into consideration the seeds (6.80) and (6.75), we can construct the matrices G22 , G23 , G24 and so on as follows: ( )( ) ( ) G = 2 ( 5 2) + ( 2 1) = (12 5) 21 10 5 2 G = 2 (12 5) + (5 2) = ( 29 12) . 5 2 21 12 5 G22 = 2 2 1 + 1 0 = 5 2 10 0 1 21 (6.81) 3 2 (6.82) 4 2 (6.83) Using the recursive relation (6.79) and taking into consideration the seeds (6.80) and (6.75), we can construct the matrices G2−1 , G2−2 , G2−3 and so on as follows: ( ) ( )( ) G = ( 1 0 ) − 2 ( 0 1 ) = ( 1 −2 ) 01 1 −2 −2 5 G = ( 0 1 ) − 2 ( 1 −2) = ( −2 5 ) . −2 5 1 −2 5 −12 G2−1 = 2 1 − 2 1 0 = 0 1 10 0 1 1 −2 (6.84) −2 2 (6.85) −3 2 (6.86) A sequence of the matrices G2n is represented in Table 6.6. THE MATHEMATICS OF HARMONY Alexey Stakhov 338 Table 6.6. A sequence of the matrices G2n n 0 1 2 1 1 0 0 1 1 −2 2 3 12 5 5 2 5 −2 5 −12 ( ) ( ) ( ) ( ( ) ( ) ( ) ( 1 0 1 0 n 2 G G2− n 0 1 0 1 5 2 1 −2 2 1 −2 5 ) ( ) ( 4 ) ( ) ( 29 12 5 −12 12 5 −12 29 5 70 29 −12 29 ) 29 12 29 −70 ) Consider the case m=3. Remember that for the case m=3 a sequence of the Fibonacci mnumbers Fm(n) looks similar to Table 6.7. Table 6.7. A sequence of the Fibonacci mnumbers for the case m=3 n 6 5 4 3 2 1 0 −1 −2 −3 −4 −5 −6 F3 ( n ) 360 109 33 10 3 1 0 1 −3 10 −33 109 −360 Using the recursive formulas (6.78) and (6.79) at the seeds (6.75) and (6.80) for the case m=3 we can construct the matrices G3n (see Table 6.8). Table 6.8. A sequence of the matrices G3n n 0 1 3 1 1 0 0 1 1 −3 2 10 3 3 1 1 −3 −3 10 ( ) ( ) ( ( ) ( ) ( 1 0 1 0 n 3 G −n 3 G 0 1 0 1 3 33 10 10 3 −3 10 10 −33 ) ( ) ( 4 109 33 33 10 10 −33 −33 109 ) ( ) ( 5 360 109 109 33 −33 109 109 −360 ) ( ) ( ) ) It is easy to verify that all square matrices of the kind Gmn in Table 6.6 and Table 6.8 possess one surprising property: all their determinants are equal to +1 (for the even powers n) or (−1) (for the odd powers n). In fact, the determinant ( ) ( ) is equal to (−33)×(−360)−109×109=−1. 29 12 of the matrix G2 = 12 5 is equal to 29×5−12×12=1 and the determinant of the 4 −33 109 matrix G3 = 109 −360 Now, let us consider a general case of m. Remember that the number m takes its value from the range of positive real numbers, for example, m can take the following values: m = 2, π, e (the base of natural logarithms) and so on. Using the recursive relation (6.78) and taking into consideration the seeds (6.75) and (6.80), we can construct the matrices Gm2 , Gm3 , Gm4 and so on for a general case of m: −5 m 1 1 0 m2 + 1 m Gm2 = m + = 1 0 0 1 m 1 (6.87) 2 m 1 m 3 + 2m m2 + 1 Gm3 = m m + 1 m + = 2 m m 1 1 0 m + 1 (6.88) Chapter 6 Fibonacci and Golden Matrices 339 m 3 + 2m m2 + 1 m2 + 1 m m 4 + 3m2 m 3 + 2m Gm4 = m 2 + = . (6.89) m m 1 m 3 + 2m m2 + 1 m +1 Using the recursive relation (6.79) and taking into consideration the seeds (6.75) and (6.80), we can construct the matrices Gm−1 , Gm−2 , Gm−3 and so on for another general case of m: ( ) ( ) ( Gm−1 = m 1 − m 1 0 = 0 1 1 0 0 1 1 −m ) (6.90) 1 −m Gm−2 = 2 −m m + 1 (6.91) −m m2 + 1 Gm−3 = 2 3 . m + 1 − m − 2m (6.92) 6.6.5. The Inverse Matrices Gmn Again, let us consider Tables 6.6 and 6.8. They set the “direct” and “inverse” Gmmatrices. Compare the “direct” and the “inverse” Gmmatrices, Gmn and Gm−n . It is easy to find a very simple method to obtain the “inverse” matrix Gm−n from its “direct” matrix Gmn . In fact, if the power n of the “direct” matrix Gmn given by (6.72) is even (n=2k), then for obtaining its inverse matrix Gm−n it is necessary to interchange the places of the diagonal elements F m(n+1) and F m(n 1) in (6.72) and to take the diagonal elements F m (n) in (6.72) with the opposite sign. This means that for the case n=2k the “inverse” matrix Gm−2 k has the following form: F (2k − 1) − Fm (2k) Gm−2 k = m (6.93) . − Fm (2k) Fm (2k + 1) To obtain the “inverse” matrix Gm−n from the “direct” matrix Gmn given by (6.72) for the case n = 2k + 1, it is necessary to interchange the places of the diagonal elements Fm(n+1) and Fm(n1) in (6.72) and to take them with the opposite sign, that is: − F (2k − 1) Fm (2k) Gm−2 k −1 = m (6.94) . Fm (2k) − Fm (2k + 1) Another way of obtaining the matrices Gmn follows directly from the ex pression (6.72). Let us consider the case m=2. In order to get the matrices G2n (n=0,±1,±2,±3,…) we have to represent two Fibonacci series F2(n+1) and F2(n) shifted by one number with respect to the other (Table 6.9). Alexey Stakhov THE MATHEMATICS OF HARMONY 340 Table 6.9. The shifted Fibonacci series F2(n+1) and F2(n) n 6 5 4 3 2 1 0 −1 −2 −3 −4 5 −6 F2 ( n + 1) 169 70 29 12 5 2 1 0 1 −2 5 12 29 F2 ( n ) 70 29 12 5 2 1 0 1 −2 5 −12 29 70 If we select the number n=1 in the first row of Table 6.9 and then select four Fibonacci numbers in the lower two rows under the number 1 and to the right with respect to it, then a totality of these Fibonacci numbers build up the Gm matrix (6.72). The Gmmatrix is singled out by bold type in Table 6.9. If we move in Table 6.9 to the left with respect to the Gmmatrix, then we obtain the matri ces G22 , G23 , G24 and so on. If we move in Table 6.9 to the right with respect to the G mmatrix, we then obtain the matrices G20 = I , G2−1 , G2−2 and so on. Note that the matrix G25 and the “inverse” to it ma trix G2−5 are singled out by bold type in Table 6.9. Note that the matrix G20 = 1 0 01 is an identity matrix. ( ) This method of obtaining the matrices Gmn can be used for any arbitrary m. 6.7. Fibonacci Qp,mMatrices 6.7.1. A Definition of the Qp,mMatrix In order to define the Qp,mmatrix, let us return back to the recursive rela tion (4.295) at the seeds (4.296) that generate the Fibonacci (p,m)numbers. By analogy with the Qpmatrix (6.28), which is a generating matrix for the Fibonac ci pnumbers Fp(n), we introduce the Qp,mmatrix (6.95), which is a generating matrix for the recursive relation (4.295) at the seeds (4.296). Given integer p>0 and real number m>0, the generating matrix for the Fibonacci (p, m)numbers is m 1 0 0 " 0 0 0 0 1 0 " 0 0 0 0 0 1 " 0 0 Q p ,m = # # # # # # # . (6.95) 0 0 0 0 " 1 0 0 0 0 0 " 0 1 1 0 0 0 " 0 0 Note that the expression (6.95) defines an infinite number of square matri ces because every positive real number m generates its own generating matrix of Chapter 6 Fibonacci and Golden Matrices 341 the kind (6.95). The Qp,mmatrix (6.95) is a square (p+1)×(p+1)matrix. It con sists of the identity (p×p)matrix bordered by the last row consisting of 0’s and the leading 1, and the first column consisting of 0’s embraced by the upper ele ment of m and the lower element of 1. We can list the Qp,mmatrices for the case p=1, 2, 3, 4 as follows: m 1 0 0 0 m 1 0 0 0 0100 m 1 0 0 0 1 0 1 m ; Q = 0 0 0 1 0. Q1,m = = Gm ; Q 2,m = 0 0 1 ; Q 3,m = 4 , m 1 0 1 0 0 0 0 0 1 0 0 0 0 1 (6.96) 1 0 0 0 1 0 0 0 0 Note that the Qp,mmatrix (6.95) is a wide generalization of the Qmatrix (6.2) (p=1, m=1), the Qpmatrix (6.28) (m=1) and Gmmatrix (6.70) (p=1). ( ) 6.7.2. The Main Theorems for the Powers of the Qp,mMatrices Now, let us raise the Qp,mmatrix (6.95) to the nth power and find the analytical expression for matrix Q pn,m . We have the following theorem for the Qp,mmatrices. Theorem 6.17. For a given integer p=1,2,3,…, we have: Fp , m (n) " Fp , m (n − p + 2) Fp, m (n − p + 1) Fp ,m (n + 1) F (n − p + 1) F (n − p) " F (n − 2 p + 2) F (n − 2 p + 1) p,m p,m p,m p,m n Q p,m = # # # # # , Fp , m (n − p − 1) Fp ,m (n − 1) Fp ,m (n − 2) " Fp , m (n − p) F ( n) Fp , m (n − 1) " Fp , m (n − p + 1) Fp , m (n − p) p,m (6.97) where Fp,m(n) is the Fibonacci (p,m)number, and n=0,±1,±2,±3,… . Proof. For a given p we can use an induction method. (a) The basis of the induction. First, we will prove the basis of the induc tion, that is, we prove that the statement (6.97) is valid for the case n=1. It is clear that for the case n=1 the matrix (6.97) takes the following form: Fp ,m (1) " Fp ,m (− p + 3) Fp ,m (− p + 2) Fp,m (2) Fp ,m (− p + 2) Fp ,m (− p + 1) " Fp ,m (−2 p + 3) Fp,m (−2 p + 2) . # # # # # Fp ,m (−1) " Fp ,m (− p + 1) F p , m ( − p) Fp,m (0) F (1) Fp ,m (0) " Fp ,m (− p + 2) Fp ,m (− p + 1) p ,m (6.98) Taking into consideration the values of the initial terms of the Fibonacci (p,m)numbers given by (4.296) and their mathematical properties given by (4.297), (4.299) and (4.300), we can write: Fp,m(2)=m, Fp,m(2)=1 and Fp,m(p)=1. THE MATHEMATICS OF HARMONY Alexey Stakhov 342 Note that all the remaining entries of the matrix (6.98) are equal to 0. This means that the matrix (6.98) coincides with the Qp,mmatrix given by (6.95). The basis of the induction is proved. (b) Inductive hypothesis. Suppose that the statement (6.97) is valid for an arbitrary n and prove it for n +1. For this purpose we consider the product of the matrices Q pn,+m1 = Q pn,m ×Q p,m m 1 0 0 " 0 0 F ( n) " F ( n − p + 2) F ( n − p + 1) 0 0 1 0 " 0 0 F ( n + 1) F ( n − p + 1) F ( n − p) " F ( n − 2 p + 2) F ( n − 2 p + 1) 0 0 0 1 " 0 0 × # # # # % # # . = # # # # # (6.99) F ( n − 1) F (n − 2) " F (n − p) F ( n − p − 1) 0 0 0 0 " 1 0 F (n) 0 0 0 0 " 0 1 F ( n − 1) " F ( n − p + 1) F (n − p ) 1 0 0 0 " 0 0 Let us consider the matrix Q pn+1 = (aij ) . In order to calculate the entry a11, we multiply term by term the first row of the matrix Q pn,m by the first column of the matrix Qp,m as follows: p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m p,m a 11=mFp,m(n+1)+Fp,m(np+1). (6.100) According to the recursive relation (4.295) we have: a11=Fp,m(n+2). Continuing the process of the matrix multiplication for the example (6.99), we obtain the following entries of the first row of the matrix Q pn,+1 m : Fp,m(n+2), Fp,m(n+1), …, Fp,m(np+3), Fp,m(np+2). (6.101) Now, let us compare the first row of the matrix Q pn,m given by (6.97) with the first row of the matrix Q pn,+1 m given by (6.101). We can see that all argu ments of the Fibonacci (p,m)numbers in the sequence (6.101) differ by +1 from the arguments of the corresponding Fibonacci (p,m)numbers, which build up the first row of the matrix (6.97). By analogy, we can show that this rule is valid for all entries of the matrix Q pn,+1 m that are formed from the matrix Q pn,m by means of its multiplication by the matrix Qp,m in accordance with (6.99). This consideration proves the validity of Theorem 6.17. Also, it is easy to prove the following identity for the product of these Q pn,m matrices: Q pn,+m1 = Q pn,m ×Q p,m = Q p,m ×Q pn,m . Using (6.102), we can prove the following theorem. Theorem 6.18. (6.102) Q pn,m ×Q pk ,m = Q pk ,m ×Q pn,m = Q pn,+mk , (6.103) where n and k are integers. Chapter 6 Fibonacci and Golden Matrices 343 Decomposing each entry of the matrix (6.97), that is, the corresponding Fi bonacci (p,m)number and taking into consideration the recursive relation (4.295), we obtain the following result given by Theorem 6.19. Theorem 6.19. Q pn,m = mQ pn−,m1 + Q pn−,mp−1 . (6.104) The recursive relation (6.104) can also be represented in the following form: Q pn−,mp−1 = Q pn,m −Q pn−,m1 . (6.105) Using (6.104) and (6.105), we can construct different sequences of the Q pn,m matrices. For example, for the case p=1 the sequences of the Q1,nm matri ces for the cases m=2 and m=3 are given by Tables 6.6 and 6.8, respectively. 6.8. Determinants of the Qp,mMatrices and their Powers 6.8.1. Determinant of the Qp,mMatrix For the case p=1 the Qp,mmatrix is reduced to the Gmmatrix, that is: ( ) Q1,m = m 1 = Gm . 1 0 (6.106) According to (6.71) we have: DetQ1,m=−1. Now, let us calculate the determinant of the Q2,mmatrix (6.107) m 1 0 Q2,m = 0 0 1 . (6.108) 1 0 0 By comparing the Q2,mmatrix (6.108) with the Q1,mmatrix (6.106), we can see that the Q2,mmatrix (6.108) is reduced to the Q1,mmatrix (6.106) if we strike out the 2nd row and the 3rd column from the Q2,mmatrix (6.108). Note that the sum of these numbers 2+3=5. According to the matrix theory [157] we have: Det Q2,m=1×(1)5Det Q1,m=1×(1)×(1)=1. (6.109) Now, let us calculate the determinant of the Q3,mmatrix m 1 0 0 Q3,m = 0 0 1 0 . 0 0 0 1 1 0 0 0 (6.110) Alexey Stakhov THE MATHEMATICS OF HARMONY 344 Comparing the matrix (6.110) with the matrix (6.108), we can see that the Q3,mmatrix (6.110) is reduced to the Q2,mmatrix (6.108) if we strike out the 3rd row and the 4th column (3+4=7) in the matrix (6.110). This means [157] that Det Q3,m=1×(−1)7×Det Q2,m=1×(−1)×1=−1. By analogy, we can show that (6.111) (6.112) Det Q4,m=1×(−1)9×Det Q3,m=1×(−1)×(−1)=1. Continuing this process, that is, generating the formulas similar to (6.109), (6.111), and (6.112), we can write the following recursive relation, which connects the determinants of the adjacent matrices Qp,m and Qp1,m : (6.113) Det Qp,m=1×(−1)2p+1×Det Qp1,m . It is clear that Det Qp,m is equal to 1 for the even values of p=2k and to (1) for the odd values of p=2k +1. We can formulate this result in the form of the following theorem. Theorem 6.20. For a given p=1,2,3,…, we have: (6.114) Det Qp,m=(−1)p. Now, let us calculate the determinants of the inverse matrices of the kind −1 Q p,m . For this purpose we consider this wellknown correlation connecting the “direct” and “inverse” matrices: Q −p,1m ×Q 1p,m = I , (6.115) where I is the identity matrix with Det I=1. Applying to (6.115) the wellknown rule that is given by Theorem 6.3, we can write: Det Q −p,1m × Det Q 1p,m = Det I . (6.116) In accordance with (6.114) the determinant of the matrix Det Qp,m takes one of the two values, 1 or 1. It follows from (6.116) that Det Q −p,1m = Det Q p,m . (6.117) 6.8.2. Determinant of the Matrix Q pn,m Let us calculate the determinant of the matrix Q pn,m given by (6.97). Us ing the property (6.1), we can write: Det Q pn,m = (Q p,m ) , (6.118) where n=0,±1,±2,±3,… . Taking into consideration the result of Theorem 6.20, we can formulate the following theorem. n Chapter 6 Fibonacci and Golden Matrices 345 Theorem 6.21. For given p=1,2,3,… and n=0,±1,±2,±3,… the determi nant of the matrix Q pn,m is given by the following formula: Det Q pn,m = (−1) pm . (6.119) 6.9. The Golden QMatrices 6.9.1. A Definition of the Golden Matrices We can represent the matrix (6.4) in the form of two matrices that are given for the even (n=2k) and odd (n=2k+1) values of n: F F Q 2 k = 2 k +1 2 k (6.120) F2 k F2 k −1 F F Q 2 k +1 = 2 k +2 2 k +1 . (6.121) F 2 k +1 F2 k In Chapter 5 we introduced the socalled Symmetric Hyperbolic Fibonacci and Lucas Functions (5.57)(5.60) that are connected with the Fibonacci and Lucas numbers by the simple correlations (5.61). Consider once again the symmetric hyperbolic Fibonacci functions: Symmetric hyperbolic Fibonacci sine τ x − τ− x sFs ( x ) = (6.122) 5 Symmetric hyperbolic Fibonacci cosine τ x + τ− x cFs ( x ) = (6.123) 5 Remember that the symmetric hyperbolic Fibonacci functions (6.122) and (6.123) are connected by the following surprising identities: sFs ( x ) − cFs ( x − 1) cFs ( x − 1) = −1; 2 (6.124) 2 cFs ( x ) − sFs ( x − 1) sFs ( x − 1) = −1. Note that the identities (6.124) are a generalization of the Cassini formu la for the continuous domain. Remember that the Fibonacci numbers Fn are connected with the func tions (6.122) and (6.123) by the simple correlation: Alexey Stakhov THE MATHEMATICS OF HARMONY 346 for n = 2k sFs ( n ) , Fn = (6.125) cFs ( n ) , for n = 2k + 1. We can use the correlation (6.125) to represent the matrices (6.120) and (6.121) in terms of the symmetric hyperbolic Fibonacci functions (6.122) and (6.123): sFs(2k) cFs(2k + 1) Q 2k = cFs(2k − 1) sFs(2k) (6.126) sFs(2k + 2) cFs(2k + 1) , (6.127) sFs(2k) cFs(2k + 1) where k is a discrete variable, k=0,±1,±2,±3,… . If we exchange the discrete variable k in the matrices (6.126) and (6.127) for the continuous variable x, we obtain two unusual matrices that are func tions of the continuous variable x: Q 2k +1 = sFs(2 x ) cFs(2 x + 1) Q 2x = cFs(2 x − 1) sFs(2 x ) (6.128) sFs(2 x + 2) cFs(2 x + 1) . (6.129) sFs(2 x ) cFs(2 x + 1) It is clear that the matrices (6.128) and (6.129) are a generalization of the Q matrix (6.4) for the continuous domain. They have a number of unique mathe matical properties. For example, for x=1/4 the matrix (6.128) takes the follow ing form: 3 1 cFs sFs 1 2 2 . Q2 = Q = (6.130) 1 1 sFs cFs − 2 2 It is difficult to even imagine what a “square root of the Qmatrix” means. However, such an amazing “Fibonacci fantasy” follows directly from (6.130). Q 2 x +1 = 6.9.2. The Inverse Golden Matrices In the above we introduced the “inverse” Fibonacci Qmatrices given by (6.66) and (6.67). We can represent the inverse of matrices (6.66) and (6.67) in terms of the symmetric hyperbolic Fibonacci functions (6.122) and (6.123) as follows: Chapter 6 Fibonacci and Golden Matrices 347 cFs(2k − 1) − sFs(2k) Q −2 k = − sFs(2k) cFs(2k + 1) (6.131) − sFs(2k) cFs(2k + 1) , (6.132) cFs(2k + 1) − sFs(2k + 2) where k is a discrete variable, k=0,±1,±2,±3,… . If we exchange the discrete variable k in the matrices (6.131) and (6.132) for the continuous variable x, we then obtain the following matrices that are functions of the continuous variable x: Q −2k−1 = cFs(2 x − 1) − sFs(2 x ) Q −2 x = − sFs(2 x ) cFs(2 x + 1) (6.133) − sFs(2 x ) cFs(2 x + 1) . (6.134) cFs(2 x + 1) − sFs(2 x + 2) It is easy to prove that the matrices (6.133) and (6.134) are the inverse of the matrices (6.128) and (6.129), respectively, that is, Q −2 x −1 = Q2x× Q2x =I and Q2x+1× Q2x1 =I , where I is an identity matrix. 6.9.3. Determinants of the Golden Matrices Now, let us calculate the determinants of the matrices (6.128) and (6.129): Det Q 2 x = cFs ( 2 x + 1) cFs ( 2 x − 1) − sFs ( 2 x ) 2 (6.135) Det Q 2 x +1 = cFs ( 2 x + 2 ) cFs ( 2 x ) − sFs ( 2x + 1) . 2 (6.136) Compare formulas (6.135) and (6.136) with the identities (5.124) for the symmetric hyperbolic Fibonacci functions. As the identities (6.124) are valid for all values of the variable x, in particular, for the value of 2x, the next identities follow from (6.135), (6.134) and (6.124): DetQ2x=1 (6.137) DetQ2x+1=1 (6.138) Note that the unusual identities (6.137) and (6.138) are a generalization of the “Cassini formula” for the continuous domain. Alexey Stakhov THE MATHEMATICS OF HARMONY 348 6.10. The Golden GmMatrices 6.10.1. A Definition of the Golden GmMatrices The Fibonacci Gmmatrices (6.72) are a source for the wide generalization of “golden” matrices introduced above. We can represent the Gmmatrix (6.72) in the form of two matrices that are given for the even (n=2k) and odd (n=2k+1) values of n: Fm (2k) F (2k + 1) Gm2 k = m Fm (2k − 1) Fm (2k) F (2k + 2) Fm (2k + 1) Gm2 k +1 = m . Fm (2k) Fm (2k + 1) Let us consider the hyperbolic Fibonacci and Lucas mfunctions. Hyperbolic Fibonacci mfunctions sFm ( x ) = cFm ( x ) = Φ mx − Φ m− x 4 + m2 Φ mx + Φ m− x 4 + m2 x −x 2 m + 4 + m2 m m + + 4 = − 2 2 4 + m 2 (6.139) (6.140) 1 (6.141) −x x 2 m + 4 + m2 4 m m + + = + . (6.142) 2 2 4 + m2 1 The Fibonacci mnumbers are determined identically by the hyperbolic Fi bonacci and Lucas mfunctions as follows: for n = 2k sF ( n ) , Fm ( n ) = m cFm ( n ) , for n = 2k + 1. (6.143) As is shown in Chapter 5, the hyperbolic Fibonacci mfunctions (6.141) and (6.142) possess the following unique properties: sFm ( x ) − cFm ( x + 1) cFm ( x − 1) = −1 (6.144) cFm ( x ) − sFm ( x + 1) sFm ( x − 1) = 1. (6.145) 2 2 Using (6.143), we can represent the matrices (6.139) and (6.140) in terms of the hyperbolic Fibonacci mfunctions (6.122) and (6.123) as follows: Chapter 6 cFm (2k + 1) Gm2k = sFm (2k) Fibonacci and Golden Matrices 349 sFm (2k) cFm (2k − 1) (6.146) sF (2k + 2) cFm (2k + 1) Gm2 k +1 = m , (6.147) sFm (2k) cFm (2k + 1) where k is a discrete variable, k=0,±1,±2,±3,… . If we exchange the discrete variable k in the matrices (6.146) and (6.147) for the continuous variable x, we then obtain two unusual matrices that are functions of the continuous variable x: sFm (2 x ) cF (2 x + 1) Gm2 x = m cFm (2 x − 1) sFm (2 x ) (6.148) sF (2 x + 2) cFm (2 x + 1) Gm2 x +1 = m . (6.149) sFm (2 x ) cFm (2 x + 1) Note that the “golden” Gmmatrices given by (6.148) and (6.149) provide a wide generalization of the “golden” Qmatrices given by (6.128) and (6.129). The “golden” Qmatrices (6.128) and (6.129) are partial cases of the matrices (6.148) and (6.149) for the case m = 1, that is, G12 x = Q 2 x and G12 x +1 = Q 2 x +1 . (6.150) 6.10.2. Inverse Golden Gmmatrices We can represent the inverse Gmmatrices (6.93) and (6.94) in terms of the hyperbolic Fibonacci mfunctions (5.105) and (5.106), that is, cF (2k − 1) − sFm (2k) Gm−2 k = m − sFm (2k) cFm (2k + 1) (6.151) cFm (2k + 1) − sFm (2k) Gm−2 k −1 = , (6.152) cFm (2k + 1) − sFm (2k + 2) where k is a discrete variable, k=0,±1,±2,±3,… . If we exchange the discrete variable k in the matrices (6.151) and (6.152) for the continuous variable x, we then obtain the following matrices, which are functions of the continuous variable x: cF (2 x − 1) − sFm (2 x ) Gm−2 x = m − sFm (2 x ) cFm (2 x + 1) (6.153) Alexey Stakhov THE MATHEMATICS OF HARMONY 350 cFm (2 x + 1) − sFm (2 x ) Gm−2 x −1 = . (6.154) cFm (2 x + 1) − sFm (2 x + 2) It is easy to prove that the matrices (6.153) and (6.154) are the inverse of the matrices (6.148) and (6.149), respectively, that is, Gm2 x ×Gm−2 x = I and Gm2 x +1 ×Gm−2 x −1 = I , where I is an identity matrix. 6.10.3. Determinants of the Golden GmMatrices Calculate the determinants of the “golden” mmatrices (6.148) and (6.149): Det Gm2 x = cFm ( 2 x + 1) × cFm ( 2 x − 1) − sFm ( 2 x ) 2 (6.155) Det Gm2 x +1 = sFm ( 2 x + 2 ) × sFm ( 2 x ) − cFm ( 2 x + 1) . 2 (6.156) Let us compare the formulas (6.155) and (6.156) with the identities (5.136) and (5.137) for the hyperbolic Fibonacci mfunctions. As the identities (5.136) and (5.137) are valid for all values of the variable x, in particular, for the value of 2x, the following identities follow from this consideration: Det Gm2 x = 1 (6.157) (6.158) Det Gm2 x +1 = −1. Note that the unusual identities (6.157) and (6.158) are further generaliza tions of the Cassini formula for the continuous domain. 6.11. The Golden Genomatrices by Sergey Petoukhov 6.11.1. The Genetic Code The discovery of the genetic code, which is general for all living organ isms from a bacterium to a man, led us to the development of the informa tional point of view for living organisms. As it is emphasized in [59], “from this point of view all organisms are informational essences. They exist because they get hereditary information from ancestors and they live to transfer the informational genetic code to descendants. Given such an approach we may treat all other physical and chemical mechanisms pre sented in living organisms as auxiliary, helping promote the realization of this basic, informational problem.” Chapter 6 Fibonacci and Golden Matrices 351 The basis for the hereditary information language is amazingly simple. For recording genetic information in the ribonucleic acids (RNA) of any organism, the “alphabet” that consists of the four “letters” of nitrogenous bases, is used: Adenine (A), Cytosine (C), Guanine (G), Uracil (U) [in DNA instead of Uracil(U) the related Thymine (T) is used]. The genetic information transferred by molecules of heredity (DNA and RNA) defines the primary structure of the protein of the living organism. Each coded protein represents a chain that consists of 20 kinds of amino acids. The block formed from three adjacent nitrogen bases is known as a Triplet. It is possible to make 43=64 triplets from the fourletter alphabet. The genetic code is called a Degenerate one because 64 triplets code only 20 amino acids. If any protein chain contains n amino acids, then the sequence of triplets that corre sponds to it contains 3n nitrogen bases in the DNA molecule or, in other words, is given by a 3ntriplet. Protein chains usually contain hundreds of amino acids and are accordingly represented by rather long polytriplets. Recently the Russian researcher Sergey Petoukhov made an original dis covery in the genetic code [59]. This discovery shows the fundamental role of the golden ratio in the genetic code. It provides further evidence that the golden ratio underlies all Living Nature! 6.11.2. Symbolic Genomatrices Petoukhov’s basic idea [59] consists of the representation of the genetic code in matrix form. A square (2×2)matrix P is an elementary matrix that is used for the representation of a system of four nitrogen bases (“letters”) of the genetic alphabet: ( ) P= C A . U G (6.159) Sergey Petoukhov further suggests that one may represent a family of genetic codes of identical length in the form of a corresponding family of matrices P(n) that are tensor (Kronecker) degrees of the initial matrix (6.159). The matrices P(n) are named Symbolic Genomatrices. For a large enough n the given family of symbolic genomatrices P(n) represents all systems of ge netic polyplets including Monoplets of the genetic alphabet (6.159) and Triplets (6.161) that are used for coding amino acids. In each of the four quadrants of the genomatrix P(n) we can see all nplets that begin from one of the letters C, A, U, G. If we ignore the first letter of the nplets, we can see that every quadrant of the matrix P(n) reproduces the matrix P(n1) of the preceding generation. On the other hand, every matrix Alexey Stakhov THE MATHEMATICS OF HARMONY 352 P(n) builds up a quadrant of the matrix P(n1)of the next generation. Thus, a genom atrix of every new generation contains in itself in latent form all the information of the preceding generations. It is pertinent here to compare the genomatrices P(n) and the Qpmatrices (6.28). Examples of the matrices P(2) and P(3) are repre sented below: CC CA AC AA ( 2) P = P ⊗ P = CU CG AU AG UC UA GC GA (6.160) UU UG GU GG CCC CCA CAC CAA ACC ACA AAC AAA CCU CCG CAU CAG ACU ACG AAU AAG CUC CUA CGC CGA AUC AUA AGC AGA CUU CUG CGU CGG AUU AUG AGU AGG ( 3) P =P⊗P⊗P = . UCC UCA UAC C UAA GCC GCA GAC GAA (6.161) UCU UCG UAU UAG GCU GCG GAU GAG UUC UUA UGC UGA GUC GUA A GGC GGA UUU UUG UGU UGG GUU GUG GGU GGG We can see that the matrices (6.159)(6.161) possess the properties sim ilar to the matrix (6.28) because any Qpmatrix (6.28) comprises the informa tion about all the previous matrices Qp1, Qp2,…, Q1, Q0. However, on the other hand, the Q pmatrix (6.28) is contained in all the next matrices Qp+1, Qp+2, Qp+3,… . 6.12.3. Numerical Genomatrices If we substitute some numerical parameter for every symbol of nitro gen bases in the symbolic genomatrix, we obtain a Numerical Genomatrix. To form such numerical parameters, Petoukhov suggests that one uses the numerical parameters of the complementary hydrogen relation for the ni trogen bases of the genetic code. We are talking about the two or three hydrogen relations that connect complementary pairs of nitrogen bases in molecules of heredity. For the bases C and G the number of such nitrogen relations is equal to 3; however, for A and U it is equal to 2. Petoukhov suggested the following rule for obtaining the numerical genomatrix from the corresponding symbolic genomatrix. Petoukhov’s rule 1. To obtain the numerical genomatrix from the corre sponding symbolic genomatrix, it is necessary to substitute every polyplet for the product of the numbers of hydrogen relations of its nitrogen bases, namely A=U=2 and C=G=3. For example, triplet CGA in the octet matrix (6.164) is replaced by the product 3×3×2=18. Chapter 6 Fibonacci and Golden Matrices 353 As a result of such substituting, the symbolic genomatrices (6.159)(6.161) are converted, respectively, into the following numerical genomatrices Pmult: ( ) (1) Pmult = 3 2 2 3 9 6 6 4 (2) Pmult = 6 9 4 6 6 4 96 4 6 6 9 (6.162) (6.163) 27 18 18 12 18 12 12 8 = 125 18 27 12 18 12 18 8 12 = 125 18 12 27 18 12 8 18 12 = 125 12 18 18 27 8 12 12 18 = 125 (3) Pmult = 18 12 12 8 27 18 18 12 = 125 12 18 8 12 18 27 12 18 = 125 . (6.164) 12 8 18 12 18 12 27 18 = 125 8 12 12 18 12 18 18 27 = 125 125 125 125 125 125 125 125 125 1000 It is easy to find a number of interesting properties of the numerical genom atrices (6.162) (6.164). First of all, all numerical genomatrices (6.162) (6.164) are symmetric with respect to both diagonals and therefore are called bisymmetric [141]. Further, the sum of numbers of each line and each column of the matrices (6.162) (6.164) are equal to 5, 52=25, and 53=125, respec tively, and the total sums of numbers in the matrices (6.162) (6.164) are equal to 10, 102=100, and 103=1000, respectively. It is proven [141] that any ( n ) possesses similar properties because each such numerical genomatrix Pmult matrix is bisymmetric. Thus, the sum of numbers of each line and each column is equal to 5n, and the total sum of numbers in the matrix is equal to 10n. Already these surprising properties of Petoukhov’s numerical genomatrices create the impression of “magic.” But Petoukhov’s discovery [59] of an im probable connection of these numerical genomatrices to the golden mean is absolutely astounding! 6.12.4. The Golden Genomatrices Next let us consider the numerical genomatrix (6.162) that corresponds to the simplest symbolic genomatrix (6.159). Consider a square matrix Φ(1), where the numbers τ = 1 + 5 2 (the −1 golden mean) and τ = −1 + 5 2 are its elements: −1 Φ (1) = τ−1 τ . τ τ ( ) ( ) (6.165) THE MATHEMATICS OF HARMONY Alexey Stakhov 354 If we square the matrix (6.165), we obtain the following matrix: −1 −1 −2 2 2 2 . Φ (1) = τ−1 τ × τ−1 τ = τ + τ τ τ τ τ 2 τ −2 + τ2 (6.166) Now, let us recall the Lucas numbers Ln (1, 3, 4, 7, 11, 18, ...) and the Binet formula for Lucas numbers. We can represent the Binet formula for Lucas numbers as follows: n Ln = τn + (−1) τ−n , (6.167) where n=0,±1,±2,±3,… . For the case n=2 the identity (6.167) takes the following form: (6.168) τ 2+τ 2=L 2=3. Taking into consideration the identity (6.168), we can represent the ma trix (6.166) as follows: ( ) = (32 32). Φ( ) 1 2 (6.169) If we compare the matrices (6.162), (6.165), and (6.169), we can see a deep mathematical connection between the numerical genomatrix (6.162) and the “golden” genomatrix (6.165), because after squaring the “golden” genomatrix (6.165) we come to the numerical genomatrix (6.169). ( n) Petoukhov proved [59] that every numerical genomatrix of the kind Pmult has the “golden” genomatrix Φ(n), which after its squaring coincides with the initial numerical genomatrix. Petoukhov suggested the following rule for obtaining the “golden” genomatrix from the corresponding symbolic genom atrix. Petoukhov’s rule 2. To obtain the “golden” genomatrix from the corre sponding symbolic genomatrix, it is necessary to substitute every polyplet for the product of the following values of its letters: C=G=τ, A=U=τ1. If we apply this rule to the symbolic genomatrix (6.160), we obtain the following “golden” genomatrix: τ 2 τ 0 τ 0 τ −2 0 2 −2 0 2 Φ ( ) = τ0 τ−2 τ 2 τ0 . τ−2 τ 0 τ0 τ2 τ τ τ τ If we square the matrix (6.170), we obtain: τ +τ +τ +τ τ + τ + τ + τ (2) Φ = τ +τ +τ +τ τ +τ +τ +τ 2 4 0 0 −4 2 2 −2 −2 τ +τ +τ +τ 2 −2 2 −2 τ +τ +τ +τ 0 0 0 τ +τ +τ +τ 0 (6.170) 2 2 −2 −2 0 4 −4 0 τ +τ +τ +τ 0 0 0 0 τ +τ +τ +τ −2 −2 τ +τ +τ +τ 2 2 −2 2 −2 0 0 0 0 0 −4 4 0 τ +τ +τ +τ −2 −2 2 2 τ 2 τ +τ +τ +τ τ +τ +τ +τ 0 0 0 τ +τ +τ +τ 0 τ +τ +τ +τ −2 −2 −4 2 −2 2 . +τ +τ +τ −2 0 2 0 2 −4 (6.171) Chapter 6 Fibonacci and Golden Matrices 355 If we use the Binet formula (6.167), we can write the following identities: (6.172) τ2+τ2=3 and τ4+τ4=7. Taking into consideration the identities (6.172), we can represent the ma trix (6.171) as follows: 9 6 6 4 Φ (2) = 6 9 4 6 . 2 6 4 9 6 4 6 6 9 After comparison of (6.163) and (6.173) we can write: 2 (2) Φ (2) = Pmult , (6.173) (6.174) that is, the “golden” genomatrix (6.170), which was obtained from the sym bolic genomatrix (6.160) by using Petoukhov’s rule 2, is converted after its squaring into the numerical genomatrix (6.173), which was obtained from the symbolic genomatrix (6.160) by using Petoukhov’s rule 1. Now, let us consider the “golden” genomatrix Φ(3), which can be obtained from the symbolic genomatrix (6.163) by using Petoukhov’s rule 2: τ3 τ1 τ1 τ −1 τ1 τ −1 τ −1 τ −3 1 3 − 1 1 − 1 1 −3 −1 τ1 τ−1 τ 3 τ1 τ −1 τ−3 τ 1 τ −1 τ τ τ τ τ τ τ τ −1 1 1 3 −3 −1 −1 τ τ τ τ τ τ τ τ1 3 Φ ( ) = 1 −1 −1 −3 3 1 1 −1 . τ τ τ τ τ τ τ τ τ −1 τ1 τ −3 τ −1 τ1 τ3 τ −1 τ1 −1 −3 1 −1 1 −1 3 1 τ τ τ τ τ τ τ τ τ −3 τ −1 τ −1 τ1 τ −1 τ1 τ1 τ3 (6.175) If we square the matrix (6.175), then after simple transformation we obtain: 2 ( 3) Φ (2) = Pmult . (6.176) It appears that a similar regularity is valid for any numerical genomatrix ( n) . This means that Sergey Petoukhov proved the following important Pmult proposition named Petoukhov’s Discovery. Petoukhov’s discovery. Let A (adenine), C (cytosine), G (guanine), and U (uracil) be nitrogen bases (“letters”) of the genetic alphabet that build up the initial symbolic matrix P = UC GA . Let P(n) be a symbolic genomatrix formed by means of a tensor (Kronecker) raising of the initial symbolic matrix P to the nth power ( ) Alexey Stakhov THE MATHEMATICS OF HARMONY 356 and let polyplets, which are built up from the “letters” A, C, G and U, be the ( n) elements of the symbolic matrix P(n). If we build a numerical genomatrix Pmult (n) from the symbolic genomatrix P by substituting for each polyplet of the matrix P(n) the products of the numbers of hydrogen relations of its nitrogen bases according to the rule: A=U=2 and C=G=3, and if we build a “golden” genomatrix Φ(n) from the symbolic genomatrix P(n) by substituting for each polyplet of the symbolic matrix P(n) the products of the following values of its “letters” according to the rule: C=G=τ and A=U=τ1 , where τ = 1 + 5 2 is the golden mean, then there is the following fundamental relation between ( n) and the “golden” genomatrix Φ(n): the numerical genomatrix Pmult ( ) 2 (n) Φ ( n ) = Pmult . This discovery of the relation between the golden mean and the genetic code allowed Petoukhov to give a new “matrixgenetic” definition based upon the golden mean. Petoukhov’s definition. The golden mean and its inverse number (τ and τ1) are the only matrix elements of the bisymmetric matrix Φ that is a root square of such bisym metric numerical genomatrix Рmult of the second order whose elements are genetic numbers of hydrogen relations (C=G=3, A=U=2). The “golden” genomatrices of Sergey Petoukhov are the most surprising application of the matrix approach to the golden section theory. Petoukhov’s discovery [59] shows a fundamental role of the “golden mean” in the genetic code. This discovery gives further evidence that the golden mean underlies all Living Nature! It is difficult to estimate the full impact of Petoukhov’s discovery for the development of modern science. It is clear that this scientif ic discovery is of equal importance to the discovery of the genetic code! We would like to end this chapter by quoting from Petoukhov’s paper [59] emphasizing the importance of the matrix approach in Fibonacci num ber and golden mean theory: “The above formulated proposition about the matrix definition and es sence of the golden section gives the possibility of considering all this mate rial and its informative interpretation from the fundamentally new, matrix point of view. The author believes that many realizations of the golden sec tion in both Organic and Inorganic Nature are connected precisely to the matrix essence and representation of the golden section. The mathematics of the golden matrices is a new branch of mathematics, which studies recur rence relations between the series of the golden matrices and also models natural systems and processes with their help.” Chapter 6 Fibonacci and Golden Matrices 357 6.12. Conclusion 1. Thus in this chapter we have developed a theory of the special class of square matrices that have exceptional mathematical properties. The begin ning of this theory is connected with the name of the American mathemati cian Verner Hoggatt – founder of the Fibonacci Association. The theory of the Fibonacci Qmatrix was first stated in Hoggatt’s book [16]. Although the name of the Qmatrix, which is a generating matrix for the classical Fi bonacci numbers, was introduced before Verner E. Hoggatt, it was from Hoggatt’s research that the idea of the Qmatrix “caught on like wildfire among Fibonacci enthusiasts. Numerous papers have appeared in ‘The Fi bonacci Quarterly’ authored by Hoggatt and/or his students and other col laborators where the Qmatrix method became a central tool in the analysis of Fibonacci properties” [158]. 2. Alexey Stakhov in his 1999 article [103] generalized the concept of the Qmatrix and introduced the Generalized Fibonacci Qpmatrix, which is a gen erating matrix for the Fibonacci pnumbers. In 2006 Stakhov introduced [118] the concept of a Gmmatrix, which is a generating matrix for the Fi bonacci mnumbers. E. Gokcen Kocer, Naim Tuglu and Alexey Stakhov in troduced [154] the concept of the Qp,mmatrix, which is a generating matrix for the Fibonacci (p,m)numbers. The “golden” Q and Gmmatrices are based on the symmetric hyperbolic Fibonacci functions and the hyperbolic Fibonac ci mfunctions. A general property of all the above matrices is the fact that their determinants are equal to +1 or 1. It is clear that a theory of similar matrices is of general mathematical interest. 3. However, the “golden” genomatrices of Sergey Petoukhov [59] are the most surprising application of the matrix approach to the golden mean theo ry. Petoukhov’s discovery [59] demonstrates the fundamental role of the “golden mean” in the genetic code. This discovery gives further evidence that the golden mean underlies all Living Nature! Petoukhov’s discovery may actually be scientifically equal in importance to the discovery of the genetic code itself! This page intentionally left blank Part III Part III. Application in Computer Science Alexey Stakhov THE MATHEMATICS OF HARMONY 360 Chapter 7 Algorithmic Measurement Theory 7.1. The Role of Measurement in the History of Science 7.1.1. What is a Measurement? In the Great Soviet Encyclopedia we can find the following definition of the measurement notion: “Measurement is an operation, by means of which the ratio of one magni tude to another homogeneous magnitude is determined; the number, which rep resents this ratio, refers to the numerical value of the subject magnitude.” A measurement is an important method of quantitative cognition of the objective world. The eminent Russian scientist Dmitry Mendeleev, the founder of the Periodic System of Chemical Elements and Father of Russian Metrology, maintained: “Science begins with a measurement. Exact science is inconceivable without a measure.” 7.1.2. The “Differentiation Principle” of “Measurement Science” The concept of a Magnitude and its Measurement belongs among the basic concepts of science. The theoretical study of this concept began developing in ancient Greece. During the 20th century the representatives of the different scientific disciplines, philosophy, physics, mathematics, information theory, psychology, economy, and so on, paid focused attention to this concept. That is why, the problem of the definition of the subject, contents and the position of measurement science in the system of contemporary sciences is of great im portance. In this Section we make a methodological analysis of the ideas, prin ciples, and scientific theories that can be united under the general name Sci ence of Measurement or Measurement Science. The existence of various directions of measurement study in modern science is a reflection of dialectic process of the Measurement Science dif Chapter 7 Algorithmic Measurement Theory 361 ferentiation as a major principle of its development. Measurement as a method of quantitative reflection of the objective properties of the Uni verse is a dialectically manysided concept. Every exact science studies measurement from its own specific point of view. It selects the measure ment property that is most important for that given science, and a study of this measurement property results in a special measurement theory. For example, in quantum physics the most essential property of quantum mea surement is the interaction between microobject and the macromeasure ment device. This problem underlies QuantumMechanical Measurement Theory [159]. In sociology, psychology, systems theory, and economics the measurement is reduced to the choice of the type of scale, to which the measurable magnitude can be referred, and therefore the Scale Problem un derlies the Psychological Measurement Theory [160]. For technical and phys ical measurement, the main problem of measurement is the choice of the System of Physical Units and the decrease of Measurement Errors. Study of these aspects of measurement resulted in the creation of Metrology the science of technical and physical measurements [161]. Measurement er rors can be considered to be “random noises” in the “measurement chan nel;” this idea underlies the Information Measurement Theory [162]. The study of measurement, as some method or algorithm to get a numerical result, then results in the Algorithmic Measurement Theory [20, 21]. 7.1.3. Applied and Fundamental Theories of Measurement As we mentioned above, the general measurement theory concerns all branches of science and technology. On the one hand, Metrology is an applied science that concerns the Engineering Sciences and Applied Physics. On the other hand, measurement concerns the fundamental problems of science, in particular, Mathematics and Theoretical Physics. Therefore, at all stages of the development of science and technology, two levels or aspects of measure ment study exist: Applied and Fundamental. The Fundamental Level assumes a study of measurement as a fundamental problem that is a source of the development of the exact sciences (in particu lar, physics, mathematics, nonphysical exact sciences); on this level the prob lem consists of revealing the most general properties and laws of measurement as the method of quantitative cognition of the objective world. Discovery of these measurement laws had a substantial influence on the development of the exact sciences. The proof of the existence of Incommensurable Seg ments made by the Pythagoreans is an example of just such a unique math Alexey Stakhov THE MATHEMATICS OF HARMONY 362 ematical discovery. This discovery determined the development of mathe matics over millennia because it resulted in Irrational Numbers, one of the most important mathematical concepts. Heisenberg’s “Uncertainty Rela tions” that underlie quantum physics fundamentally limit the exactness of quantummechanical measurement and may be considered to be the out standing physical idea in the field of the measurement. The Applied Level assumes the study of measurement from the point of view of the practical, applied problems that appear in technology and applied physics. The problem of the creation of Systems of Measurement Units (metric system, Gauss’ absolute system, etc.) that runs throughout the history of science and technology is the most typical example of the applied measurement problem. 7.1.4. What is the Fundamental Distinction between Physics and Mathematics? Physics and mathematics education programs do not, as a rule, give much attention to this important question. Its answer lies, for example, in the works of the famous physicist Brilluen [163]. By pondering mathematical theorems and physical theories, Brilluen defines the essence of the distinction as follows. While mathematics begins with definitions of dimensionless points, infinitely thin curves, and continuous spacetime, modern physics denies any real sense to such definitions. Brilluen noted that “for physicists the irrational numbers do not have any significance. It is assumed that the irrational numbers need to be defined with infinite precision because the exactness of say one in a hundred, one in a million or even one in a billion is insufficient for their definition. For the physical experiment, this does not make any sense; we can measure within the fifth or tenth digits of a decimal fraction, but there is no experiment, which would give twenty decimal digits. There is no sense in asking, whether the speed of light is a rational number, whether the ratio M/m (the ratio of the proton weight M to the electron weight m) is rational. All irrational numbers, like π, e, 2 , etc., only follow from abstract mathematical definitions. However, as we have already said, mathematics is an aimless art. No physical objects can show real geometrical properties in those borders, which are assumed by mathematicians; we can even name such an approach, “wishful thinking” [163]. Thus, according to Brilluen, a feature of the “physical” approach to the study of the objective world consists in the postulation of some limit of “definiteness” of the “physical” magnitude that is completely ignored in the purely mathemat ical approach to the definition of the concept of magnitude. Chapter 7 Algorithmic Measurement Theory 363 7.1.5. A Thesis about Inevitability of Measurement Errors Hence, according to Brilluen, the distinction between physical and mathe matical approaches to measurement theory consists of different attitudes toward the Thesis about the Inevitability of Measurement Errors. The physical approach recognizes this thesis as the main postulate of the physical measurement theory. This axiom has an empirical, practical substantiation. The Russian mechanic and academician Krylov noted that “every measurement of any magnitude al ways has some error. Clearly, the less the error, the more exact the measure ment, but the practice of measurement shows that it is impossible to avoid error entirely. This is confirmed by the fact that when iterating the same measure ment many times by the same device we will get measurement results, which are different between themselves” [161]. This axiom finds its theoretical acknowledgement in the fact of the existence of some limit of definiteness (negentropy) of physical magnitude, which is known in quantum mechanics as Heisenberg’s Uncertainty Relation, in wave mechanics as a Relation of TimeandFrequency Uncertainty, and on the molecular level is given by the laws of thermodynamics (Brilluen’s Negentropy Principle of Infor mation). The existence of the limit of definiteness of physical magnitude and, as a consequence, the impossibility of the “absolutely exact” comparison (distinc tion) of the sizes of two physical magnitudes, which differ from each other in the size of the definiteness limit, is the basic thesis of “physical” measurements. Thus, the concept of “magnitude” in physics differs from a similar concept in mathe matics, where we adopt the opposite thesis about the possibility of the “abso lutely exact” comparison of two magnitudes independently of their sizes (com parison axioms of the mathematical theory of magnitudes). 7.1.6. Purposes of the “Physical” Measurement Theory If we follow Brilluen’s approach, we can divide the “fundamental” part of measurement theory into the Physical and Mathematical measurement theo ries. This approach allows us to formulate the basic purpose of the “physical” measurement theory which follows from the basic thesis about the inevitability of measurement errors. In quantum physics [159], the main purpose of mea surement theory is the study of the interaction of the quantummechanical ob ject with the macrodevice used for the quantummechanical measurement. In metrology and the applied physics, the main purpose of the measurement theo ry is to study and model the errors of the physical measurements and in recent decades to study the sensitivity threshold of measurement devices. Of course, Alexey Stakhov THE MATHEMATICS OF HARMONY 364 the doctrine about physical magnitudes and units [161] always was an impor tant branch of the “physical” measurement theory. Study of the character of physical measurement errors resulted in the formu lation of two empirical axioms, which underlie the theory of random errors [161]: 1. Randomness’ axioms: for the very big number of measurement re sults, the random errors, equal in terms of absolute value, but varying by sign, meet equally often; the number of the negative errors is equal to the number of the positive errors. 2. Distribution’s axiom: smaller errors meet more often, than bigger; very big errors do not meet. These empirical axioms resulted in the wellknown Normal Law of Probabil ity Distribution (Gauss’ law): 2 x − 1 ϕ(x ) = e 2 σ2 , σ 2π 2 where σ is a dispersion of probabilities; x is current value of error. In the opinion of Russian mathematical authority Tutubalin [164], “the ‘Nor mal Law’ is some kind of “miracle” of probability theory, without which this theory almost would not have the original contents ....” The correctness of the above axioms and the “normal law,” which follows from them, is confirmed by the fact that all conclusions, based on them, are always consistent with experience. However, the nonstrict character of the formulated axioms always causes a certain dissatisfaction with Gauss’ law. This situation is expressed by one witty mathematician who sarcastically noted that “the experimenters believe in Gauss’ law, by relying on the proofs of mathematicians, and the mathematicians believe in this law, by relying on the experimental verification.” 7.2. Mathematical Measurement Theory 7.2.1. Evolution of the Measurement Concept in Mathematics A measurement problem plays the same exceptional role in mathematics, as in other areas of science, particularly in technology, physics, and other exact sciences. The Russian mathematician and academician Kolmogorov in the foreword to the Russian edition of Lebesgue’s The Measurement of the Magnitudes [3] expressed his outlook on this problem in the following words: “Where is the main interest of Chapter 7 Algorithmic Measurement Theory 365 Lebesgue’s book? It seems to me, it is in the following: mathematicians have a propensity already owning the finished mathematical theory to be ashamed of its origin. In comparison with the crystal clearness of the development of a theory, beginning from its already prepared basic concepts and assumptions, it seems dirty and unpleasant to delve into the origin of these basic concepts and assumptions. The entire edifice of school algebra and all mathematical analysis can be construct ed solely on the concept of real number without any reference to the measurement of concrete magnitudes (lengths, areas, time intervals, etc.). Therefore, on the different steps of education with a different degree of boldness, one and the same tendency emerges: as soon as it is possible to be done with the introduction of number concept and immediately to start discussing only numbers and relations between them. This is the tendency Lebesgue protests against this!” Let us track now the evolution of the measurement concept in mathemat ics. It is well attested that, the measurement rules used by the Egyptian land surveyors were the first “measurement theory.” The ancient Greeks testify that geometry was obliged by its origin (and the name) to those measurement rules, that is, to the “measurement problem.” However, in Ancient Greece the separation of the measurement problems into two parts began – the applied problems related to “logistics,” which was at that time referred to as a set of certain rules for the applied measurements and calculations; and fundamental problems related to geometry and number theory. These latter fundamental measurement problems became the main problems of Greek mathematics. In the Greek period, a science of measurement was developing primarily as a mathematical theory. During this period, the Greek mathematicians made the following mathematical discoveries: Incommensurable Segments, Eudoxus’ Meth od of Exhaustion and the Measurement Axiom. These outstanding mathematical discoveries later became the sources of number theory, integration and differen tiation and other mathematical theories. These mathematical discoveries had a direct impact upon the “measurement problem.” It gave the Bulgarian mathema tician and academician Iliev the impetus to proclaim that “during the first epoch of mathematics development, from antiquity to the discovery of differential and integral calculus, investigating first of all the measurement problems, mathemat ics had created Euclidean geometry and number theory” [5]. 7.2.2. Incommensurable Line Segments From high school on we believe in the strictness and stability of mathemat ics. Therefore, for some of us it is a big surprise that in the process of its develop ment, mathematics was experiencing different crises. Moreover, it may be a big Alexey Stakhov THE MATHEMATICS OF HARMONY 366 ger surprise for others that, since the beginning of the 20th century, modern mathematics is in a deep crisis, and contemporary mathematicians cannot see any way out of this crisis yet. The first crisis in the foundations of mathematics was connected directly to the “measurement problem.” The early Pythagorean mathematics was based upon the socalled Commensurability Principle. Let us recall, that in mathematics, two nonzero real numbers a and b are said to be Commensurable if and only if a/b is a rational number. According to the commensurability principle, any two geometric magnitudes Q and V have a common measure, that is, both magni tudes are divisible by it. Thus, their ratio can be expressed as the ratio of two mutually prime natural numbers m and n: Q m = . (7.1) V n According to the main Pythagorean philosophical doctrine that Everything is a number, all geometrical magnitudes can be expressed in the form (7.1), that is, their ratio is always rational. This means that geometry is reduced to a num ber theory because according to (7.1) all geometric relations can be expressed by rational numbers. Let us examine the ratio of the diagonal b and the side a of the square (Fig. 7.1). According to Pythagorean Theorem (7.2) b2=2a2. It follows from (7.2) that the ratio of the diagonal b to the side a of the square (Fig. 7.1) is equal (7.3) b / a = 2. Suppose that the diagonal b and the side a of the square are commensura ble. This means that we can represent their ratio in the form 2 = m n , where m and n are the mutually prime natural numbers. Then m2=2n2. Hence, it follows from here that the number m2 is an even number. However, if a square of a number is even, this means that the number m is also an even number. As the numbers m and n are mutually prime natural numbers, then the number n is an odd number, according to the definition (7.1). Howev a er, if m is even number, the number m2 can be divided by 4, and hence, n2 is an even number. Thus, n is also an b b even number. However, n cannot be an even and odd 2 a number simultaneously! This contradiction shows that our premise about the commensurability of the diago Figure 7.1. Incommensura nal and the side of the square is wrong and therefore the number 2 is irrational. ble line segments. Chapter 7 Algorithmic Measurement Theory 367 The discovery of incommensurability staggered the Pythagoreans and caused the first crisis in the basis of mathematics, as this discovery refuted the initial Pythagorean doctrine about commensurability of all geometric magnitudes. A discovery of irrational numbers originated a complicated mathematical notion, which did not have direct connection with human experience. According to the legend, Pythagoras made a “hecatomb,” that is, sacrificed one hundred oxen to the gods in honor of this discovery. However, according to the Brewer’s Dictio nary, “He sacrificed to the gods millet and honeycomb, but not animals. He forbade his disciples to sacrifice oxen.” In any case, it was a worthwhile sacrifice because this discovery became the turning point in the development of mathe matics. It ruined the former system created by the Pythagoreans (the commen surability principle) and became a source of new and remarkable theories. The importance of this discovery may be compared with the discovery of nonEuclid ean geometry in the 19th century or to the theory of relativity at the beginning of 20th century. Similar to these theories, the problem of incommensurable line segments was wellknown among educated people. Plato and Aristotle were also addressing the problem of incommensurability in their works. 7.2.3. EudoxusArchimedes and Cantor’s Axioms To overcome the first crisis in the mathematics foundations, the Great Greek mathematician Eudoxus developed the method of exhaustion and created a new theory of magnitudes. Eudoxus of Cnidus (c.410/408 – 355/347 BC) was a Greek astronomer, mathematician, physicist, scholar and friend of Plato. Unfortunately, all of his own works were lost, therefore we obtained information about him from oth er sources, such as Aratus’ poem on astronomy. Archytas from Athens was his mathematics teacher. He is famous due to the introduction of the astronomical globe, and his early contributions to the explanation of the planetary motion. Eudoxus developed the method of exhaustion and used it for the creation of the theory of irrationals. Later this method was used masterfully by Archimedes. The essence of Eudoxus’ method of exhaustion can be explained by the following practical example. If we have a barrel of beer and a beer mug, then this beer in the barrel will eventually be exhausted, even though the beer barrel is enormous and the beer mug very small. Eudoxus’ theory of incommensurability (see Book V of Euclid’s Ele ments) could be considered one of the greatest achievements of mathe matics and in general coincides with the modern theory of irrational num bers suggested by Dedekind in 1872. THE MATHEMATICS OF HARMONY Alexey Stakhov 368 The measurement theory of geometric magnitudes dates back to the incom mensurable line segments. It is based on the group of socalled continuity axi oms, which comprise both EudoxusArchimedes’ axiom and Cantor’s axiom (or Dedekind’s axiom). EudoxusArchimedes’ axiom (the axiom of measurement). For any two segments A and B it is possible to specify such natural number n, which always results in the following nonequality: nB>A. (7.4) In the 19th century, Dedekind and then Cantor made a final attempt to create a general theory of real numbers. For this purpose, they introduced the additional axioms into the group of the continuity axioms. For instance, let us consider Cantor’s Axiom. Cantor’s continuity axiom (Cantor’s principle of nested segments). If an infinite sequence of segments is given on a straight line A 0B0, A 1B 1, A2B2,…,AnBn, …, such that each next segment is nested within the preceding one, and the length of the segments tends to zero, then there exists a unique point which belongs to all the segments. A A0 A1 A2 An Bn B2 B1 B0 C B B B B a) B B b) Figure 7.2. The continuity axioms: (a) EudoxusArchimedus’ axiom (b) Cantor’s axiom The main result of the mathematical measurement theory that is based on the continuity axioms is a proof of the existence and uniqueness of the solution q of the Basic Measurement Equality: Q=qV, (7.5) where V is a measurement unit, Q is a measurable segment, and q is any real number named a Result of Measurement. The idea of the proof of the equality (7.5) consists of the following. Using Eu doxusArchimedes’ axiom and certain rules called a Measurement Algorithm, we can form from the measurement unit V some sequence of the nested segments, which can be compared with the measurable segment Q during the process of mea surement. If we direct this process ad infinitum, then according to Cantor’s axiom for the given line segments Q and V we always can find such nested segments, which coincides with the measurable segment Q. It follows from this proof that the idea about a measurement is one of a process running during an infinite time. Chapter 7 Algorithmic Measurement Theory 369 It is difficult to imagine that the continuity axioms and the mathematical measurement theory are the result of a more than 2000year period in the devel opment of mathematics. The continuity axioms and basic measurement equa tion (7.5) comprise in themselves a number of the important mathematical ideas that underlie different branches of mathematics. It is necessary to note that the measurement axiom expressed by (7.4) is a reflection of Eudoxus’ method of exhaustion in modern mathematics. The axiom generalizes mankind’s millennia of experience regarding the measurement of dis tances, areas and time intervals. The simplest measurement algorithm of the line segment A by using the line segment B less than A underlies this axiom. This algorithm consists of the successive comparison of the line segment B+B+…+B with the measurable line segment A. During the measurement process we count a number of the line segments B that are laid out on the measurable line segment A. This measurement algorithm is called a Counting Algorithm. The counting algorithm is the source of the various fundamental notions of arithmetic and number theory, in particular, the notions of Natural Number, the Prime and Composite numbers, as well as the notion of Multiplication, Division, etc. In this connection the Euclidean definition of prime (the “first” number) and com posite numbers (“the first number is measured only by 1,” “the composite number is measured by some number”) is of great interest. The measurement axiom origi nates a Divisibility Theorem, which plays a fundamental role in number theory. A Theory of Divisibility and a Comparison Theory are based on the Divisibility Theo rem. Thus, the simplest measurement algorithm, a Counting Algorithm, originates natural numbers and all concepts and theories connected with them. 7.2.4. A Contradiction in the Continuity Axioms The idea about measurement as a process running during infinite time is a brilliant example of the Cantorian style of mathematical thinking based on the concept of Actual Infinity. However, this concept was subjected to sharp criticism from the side of the representatives of constructive mathematics. For the explanation of the distinctions between different interpretations of infinity used in EudoxusArchimedes’ and Cantor’s axioms, once again we return back to them. In Cantor’s axiom all infinite sets of the nested segments are considered as the set given by all its objects simultaneously (Cantor’s actu al infinity). EudoxusArchimedes’ axiom, which has an “empirical origin,” is an example of a constructive axiom and in the implicit form relies on the abstrac tion of Potential Practicability because after each step it assumes the possibility of constructing the next line segment bigger than the previous one. The number Alexey Stakhov THE MATHEMATICS OF HARMONY 370 of measurement steps in EudoxusArchimedes’ axiom, necessary for realization of the condition nB >A, is always finite, but potentially with no limitation. Here we can see the example of the constructive understanding of infinity as a potential category. In this connection we can pay attention to the internal contradiction of the classical mathematical measurement theory (and as a consequence, of the the ory of real numbers), which assumes in the initial suppositions (the continuity axioms) the coexistence of two opposite ideas about infinity, that is, the actual or completed infinity in Cantor’s axiom and the potential or uncompleted infinity in the EudoxusArchimedes’ axiom. 7.3. Evolution of the Infinity Concept 7.3.1. An Infinity Concept The book Philosophy of Mathematics and Natural Science (1927, German) by the famous Germany mathematician Hermann Weyl (18851955) begins with the following words: “Mathematics is the science of infinity” [166]. Without a doubt, the infinity concept permeates throughout all mathematics, because math ematical objects are, as a rule, members of the classes or sets containing an un countable set of elements of the same kind; examples include the set of natural numbers, the set of real numbers, the set of triangles, etc. That is why the infin ity concept is necessary under a strict analysis. Starting a discussion of the infinity concept, we should say a little about the concept of the “finite” that is opposite to the notion of “infinity.” Our intellect determines that infinity is something that does not have an end, and the finite is something that does have an end. At first sight, the concept of the finite seems to us clear and selfobvious. However, in actuality it is not so simple. If we consider the finite as a selfobvious concept, then infinity can be seen as an unlimited accumulation of the finite. On the other hand, recognizing the infinite division of the finite object (for example, the geometric line seg ment that consists of points), we are compelled to state that the infinite is made from the finite and the finite is made from the infinite. The secret of the infinity in space and of the “immortality” in time has always been the most highpowered catalyst that directs the human intellect to the search for truth. David Hilbert, one of the most brilliant mathematicians of the first half of the 20th century, said: Chapter 7 Algorithmic Measurement Theory 371 “No other question has ever moved so profoundly the spirit of man; no other idea has so fruitfully stimulated his intellect; yet no other concept stands in great er need of clarification than that of the infinite” [quoted from E. Manor, To In finity and Beyond, Boston, 1987]. 7.3.2. The Origin of the Symbol ∞ The symbol for “infinity” in the form of the “horizontal 8” was introduced for the first time in Arithmetic of Infinite Values by English mathematician John Vallis in 1665. Some historians of mathematics give the following explanation for the use of this infinity symbol: a similar symbol was used in Roman notation and was designated as a thousand which was identified as “very much.” The Rus sian historian of mathematics Gleiser in his book [150] gives, however, another explanation of the origin of this symbol in Vallis’ book. According to his opin ion, the symbol ∞ can be considered as two zeros connected among themselves. This symbol was opposed by Vallis to the symbol “zero” (0) in the connection with the following expressions given in Vallis’ book: 1 / 0 = ∞, 1 / ∞ = 0. (7.6) However, at the present moment, the expressions (7.6) are considered to be an error. People should know and remember that the symbol ∞ does not mean any number and does not have any numerical sense. Therefore, it is necessary to use the symbols +∞ and −∞ with big caution; in particular, we cannot fulfill any arithmetical operations similar to (7.6) with them. 7.3.3. Potential and Actual Infinity Although according to Hermann Weyl, “infinity” is a fundamental concept of mathematics, there is nevertheless no comprehensible definition of this important concept in mathematics. Arithmetical and Geometric, Potential and Actual infini ty are used in mathematics. Let us consider these concepts in greater detail. A sequence of natural numbers 1,2,3,… (7.7) is the first and most important example of Arithmetical Infinity. Since Hegel, the arithmetical infinity of natural numbers 1+1+1+ … i.e. the endless “itera tion of the same” is called the “bad infinity.” Geometric Infinity, for instance, consists of the unlimited bisection of a line segment. Pascal wrote about geometric infinity as follows: “There is no geometer who would not suppose, that any space can be divided ad infinitum. It is impossi ble to be devoid of this, similar to a person, who cannot be without a soul. And Alexey Stakhov THE MATHEMATICS OF HARMONY 372 nevertheless there is no person who would understand that means an infinite di visibility” [166]. Except for the distinction between arithmetical and geometric infinity, there is a distinction between Actual and Potential infinity. For consid eration of the difference between these concepts, once again, we address to the sequence of natural numbers (7.7). We can suppose this sequence as if “complet ed” sequence, that is, determined by all its members simultaneously. Such pre sentation about “infinity” is named Actual Infinity. However, the sequence of natural numbers (7.7) can also be considered as the “developing” sequence that is generated according to the principle n′=n+1. (7.8) This means that each natural number can be obtained from the previous one by means of summation of 1. Such presentation of the “infinity” is named Potential Infinity. 7.3.4. An Origin and Application of the Infinity Idea in the Ancient Greek Mathematics Mathematics was turned into a deductive science in Ancient Greece. Many historians of mathematics believe that Greek mathematicians introduced into mathematics for the first time the concept of Potential Infinity. “Zeno’s paradoxes” played a major role in the development of the infinity concept. They demonstrated so well the logical difficulty of the hypotheses about the infinite division of geometric line segments and time intervals. Let us con sider two of them, Dichotomy and Achilles and the Tortoise. 1. Dichotomy (“bisection” in Greek). In this antinomy Zeno asserts that movement is impossible. Really, if a solid moves from point A to point B, this solid must first traverse the line segments in 1/2, and before that 1/4, 1/ 8, 1/16 ... of the distance between points A and B. However, the sequence of such line segments is infinite. This means that one can never leave point A! The paradox, which results in an insuperable logical impasse, is based on the theorem that the sum of the infinite set of addends is finite. 2. Achilles and the Tortoise. Zeno asserts: “The fastmoving Achilles can never catch up to the tortoise.” The proof comes to the following: let Achilles be n times faster than the tortoise and let the initial distance between them be equal to d. While Achilles overcomes this new distance, simultaneously with him the tortoise can move forward on the distance d/n; while Achilles over comes new distance, the tortoise can move forward on the new distance d/n2 and so on. Thus, the distance between Achilles and the tortoise will always be more than 0, that is, Achilles never can catch up to the tortoise. Chapter 7 Algorithmic Measurement Theory 373 Aristotle paid a great deal of attention to the concept of mathematical infin ity. He strongly objected to the use of actual infinity in mathematics. The fol lowing wellknown thesis is attributed to Aristotle: “Infinitum Actu Non Da tur;” translated from Latin this means that the existence of actuallyinfinite ob jects in mathematics is impossible. 7.3.5. Cantor’s Theory of Infinite Sets The German mathematician George Cantor (18451918) is considered to be the prime disrupter of tranquility in 19th century mathematics. The history of set theory is quite different from the history of other areas of mathematics. For most areas, we can outline a long process, in which the ideas are evolving until the ultimate flash of inspiration produces a discovery of major importance, often in many mathematicians somewhat synchronously. That nonEuclidean geome try was discovered by Lobachevski and Bolyai independently at the same time is a brilliant example of such synchronous inspiration. As for set theory, it was the creation of one person, George Cantor. Cantor’s main idea is the study of infinite sets as actualinfinite sets. The idea of onetoone correspondence between the elements of the comparable sets was used by Cantor in his research of infinite sets. If we can establish such correspondence between the elements of two sets, we can say that the sets have the same Cardinality, that is, they are equivalent. Cantor wrote “In the case of the finite sets the cardinality coincides with the number of the elements.” That is why, the cardinality is also named the cardinal (quantitative) number of the given set. This approach resulted in many paradoxical conclusions being in sharp contradiction to one’s intuition. So, in contrast to the finite sets that comply with the Euclidean axiom, “The whole is more than its parts,” the infinite sets do not comply with this axiom. For example, it is easy to prove that the set of natural numbers and some of its subsets are equivalent. In particular, we can establish the following onetoone correspondence between the sets of natural numbers and the even numbers: 1 2 3 ... n ... 7 7 7 7 . 2 4 6 ... 2n ... This feature of the infinite sets can be used for their definition: a set is named infinite if the set is equivalent in extant to one of its subsets. A set is finite if the set is not equivalent to any of its subsets. Any set equivalent to the set of natural numbers is named Denumerable because all of its elements can be numbered. Alexey Stakhov THE MATHEMATICS OF HARMONY 374 The most amazing discovery was made by Cantor in 1873. He proved that all three of the most characteristic sets (natural numbers, rational numbers, and algebraic numbers) have one and the same cardinality, that is, the sets of ratio nal and algebraic numbers are denumerable. Cantor also proved that the cardi nality of real numbers is greater than the cardinality of natural numbers. From the above consideration it is possible to conclude that in contrast to the majority of his predecessors, Cantor had undertaken for the first time a deep re search into mathematical infinity that resulted in new and unexpected outcomes. 7.3.6. Antinomies of Cantor’s Theory of Infinite Sets It seemed for many mathematicians, that Cantor’s theory of infinite sets pro duced a revolution in mathematics. The end of the 19th century culminated in the recognition of Cantor’s theory of infinite sets. French mathematician Jacques Hadamard in his 1897 speech at the First International Congress of Mathemati cians in Zurich officially proclaimed the settheoretical ideas as the basis of math ematics. In his lecture, Hadamard emphasized that the most attractive reason of Cantor’s set theory consists of supporting the fact that for the first time in math ematics history the classification of sets was made on the basis of a new concept of “cardinality” and the amazing mathematical outcomes inspired mathemati cians to new and surprising discoveries. However, very soon the “mathematical paradise” based on Cantor’s set theory was destroyed. In the first few years after the First International Congress of Mathematicians, the paradoxes or antinomies appeared in set theory. Cantor discovered the first paradoxes at the end of the 19th century. Other paradoxes were discovered later by other scientists. These paradoxes be came the basis for the next, or third (after the discovery of incommensurable line segments and substantiation of the limit theory) crisis in the foundation of mathematics. One of them, discovered by the English philosopher and mathe matician Bertrand Russell in 1902, was connected with the foundations of set theory and the concept of the set of all sets. Each of the usual sets, which we met until now, does not contain itself as an element; so, for example, the set of all natural numbers is not a natural number. We name such sets Ordinary Sets. However, there are also such exotic sets that contain themselves as their own elements, for example, the set of all sets. We name such sets Nonordinary Sets. Consider the set S of all ordinary sets and ask the following question: is S an element of S? Alternatively, is the set S ordinary or nonordinary set? Suppose that S is an element of S; then it is necessary to recognize that S is a nonordinary set. However, as S contains only ordinary sets, this means that S should also be Chapter 7 Algorithmic Measurement Theory 375 an ordinary set. Thus, from this supposition it follows that S is an ordinary set, that is, we have a logical contradiction. Now, suppose that S is not an element of S, that is, S is the ordinary set; then S should be in S, as all ordinary sets. Thus, from the supposition that S is the ordinary set it follows that S is the nonordi nary set; again, we have contradiction! Russell demonstrated this contradiction, known as the Barber Paradox, with the example of the village barber, who gave a promise to shave those and only those village inhabitants, who do not shave themselves. We can ask the ques tion: Does the barber shave himself? If he shaves himself, this means that there by he includes himself into the set of inhabitants, who shave themselves, and therefore he should not shave himself; if he does not shave himself, this means that he belongs already to those, who do not shave themselves, that is, he should shave himself. We again have a logical contradiction which of course is unac ceptable in mathematics! 7.3.7. Is Cantor’s Actual Infinity the Greatest Mathematical Mystification? A discovery of the antinomies in Cantor’s set theory caused a new crisis in the foundations of mathematics. Various attempts were made to overcome this crisis including Constructive Analysis in mathematics the most radical of them. The representatives of constructive analysis saw the main reason for paradoxes in Cantor’s theory of sets to be the use of the concept of “actual infinity.” Russian mathematician Markov, one of the brightest representa tives of constructive analysis, wrote: “It is impossible to imagine an endless process as a completed process without rough violence to our intellect, in the rejection of such inconsistent fantasies.” [167] On the other hand, probably the works of Russian mathematician Alex ander Zenkin took the final step in the dispute over Cantor’s set theory. Zenkin found logical contradictions and errors in Cantor’s set theory. The essence of Zenkin’s approach, presented in a few main publications [168], consists of the fact that for the first time he formulated an almost obvious fact – the concept of potential infinity (PI) in the Aristotelian form – “all infinite sets are poten tial sets,” and the concept of actual infinity (AI) in Cantor’s form – “all infinite sets are actual sets” are AXIOMS, that is, they cannot be proved or refuted, but they can be either accepted, or rejected. For this reason, any discussions about “existence” or “nonexistence” of the “actual infinity,” which goes back to Aristotle, cannot convince anyone of anything. Here we can provide an analogy with the history of the 5th Euclidean axiom about parallel lines. Howev Alexey Stakhov THE MATHEMATICS OF HARMONY 376 er, there is also an essential difference – the correctness of the nonEuclidean ge ometry was proven – however, the correctness of Cantor’s theory of sets based on “actual infinity” is not proven. Moreover, the antinomies of Cantor’s theory of sets demonstrate that this theory is self contradictory. After a thorough analysis of Cantor’s continuum theorem, in which Zen kin gave the logical substantiation for legitimacy of the use of the “actual infinity” in mathematics, he derived the following unusual conclusion: 1. Cantor’s proof of this theorem is neither a mathematical proof in Hilbert’s sense nor in the sense of classical mathematics. 2. Cantor’s conclusion about nondenumerability of the continuum is a jump through a potentially infinite stage, that is, Cantor’s reasoning contains the fatal logical error of “unproved basis” (a leap to a “wishful conclusion”). 3. Cantor’s theorem does, in fact, strictly mathematically prove the potential, that is, the unfinished character of the infinity of the set of all “real numbers.” Thus, Cantor proves mathematically the fundamental principle of classical logic and mathematics: “Infinitum Actu Non Datur” (Aristotle). According to the opinion of Alexander Zenkin [168], Aristotle’s famous the sis “Infinitum Actu Non Datur” – that is, the assertion about the impossibility of the existence of actuallyinfinite objects – had been supported during the last 2300 years by Aristotle’s great followers Leibniz, Cauchy, Gauss, Kronecker, Poincare, Brouwer, Vail, Lusin and by many other famous founders of classical logic and modern classical mathematics! Every one of them professionally stud ied the problem of mathematical infinity, and there were not be any doubts, that they understood the true nature of the infinity in general no worse than Cantor did. Especially, if we take into consideration the important fact, that infinity does not depend on progress of technological equipment of science, infinity nev er was and never becomes an object of the technological research, because no contemporary computers, never, by definition, would ever be able to finish an enumeration of all elements of natural series 1, 2, 3, … . For this reason – Alexander Zenkin asserted – all discussions about the pos sibility or impossibility of actual infinity during two millennia down to Cantor had only a speculative character. The infinity does not depend on the progress of science and technology! Alexander Zenkin wrote [168]: “Why was such an analysis of Cantor’s theorem not executed in time, i.e. at the end of the 19th century? It is a very crucial topic for fundamental research in the field of the psychology of scien tific knowledge.” Thus, there is the impression that the long history (starting from Greek science) of the study of the infinity concept as one of the fundamental mathe Chapter 7 Algorithmic Measurement Theory 377 matical concepts is approaching its completion. The antinomies of Cantor’s set theory, which caused the contemporary crisis in the foundations of mathematics, showed that the concept of actual infinity could not be a reliable basis for mathe matical reasoning, because from the point of view of the representatives of con structive analysis the “actual infinity” concept is internally selfcontradictory. On the other hand, the research of Russian mathematician Alexander Zenkin [168] testifies the existence of logical errors in Cantor’s theorems, and this fact gives one the right to doubt the accuracy of Cantor set theory. Thus, the concept of potential infinity introduced in the Greek mathe matics can be the only real base for mathematical reasoning. In this connec tion Aristotle’s thesis “Infinitum Actu Non Datur” should become the main axiom for the creation of new, constructive mathematics. 7.4. A Constructive Approach to Measurement Theory 7.4.1. The Rejection of Cantor’s Axiom from Mathematical Measurement Theory As the concept of “actual infinity” is an internally contradictory notion (“the completed infinity”), this concept cannot be a reasonable basis for the creation of constructive mathematical measurement theory. If we reject Cantor’s axiom, we can try to construct mathematical measurement theory on the basis of the idea of potential infinity, which underlies the EudoxusArchimedes’ axiom. By referring to the measurement theory, this means, that the number of steps for the given measurement is always finite, but potentially unlimited. The accep tance of the given approach at once results in the occurrence of the fundamental ly irremovable measurement error called Quantization Error. The constructive approach to measurement theory results in a change of the purpose of measurement theory. One of the essential moments at the proof of the equality (7.5) is a choice of the Measurement Algorithm, by means of which we carry out the formation of “nested segments” from the measurement unit V. With the actually infinite time, the measurement algorithm does not influence the measurement result q and its choice is arbitrary. With the given time of measurement – that is defined by the number of measurement steps n – and with the other equal conditions, the distinction between the measure ment algorithms with respect to the “reached” measurement exactness appears. Alexey Stakhov THE MATHEMATICS OF HARMONY 378 With these conditions, the second constructive idea about the “efficiency” of the measurement algorithms comes into effect and the Problem of the Synthe sis of the Optimal Measurement Algorithms is put forward as a central problem of the Constructive (Algorithmic) Measurement Theory. 7.4.2. BashetMendeleev’s Problem The constructive approach, at once, results in the optimization problem in measurement theory which was not a topical problem for classical measurement theory. By studying the optimization problem in the measurement theory, we come unexpectedly to a combinatorial problem known as a Problem about the Choice of the Best Weights System. For the first time, this prob lem appeared in the 1202 in Fibonacci’s Liber Abaci. From Fi bonacci’s work this problem moved to the 1494 book Summa de Arithmetica, Geomeytria, Proprtioni et Proportionalita by Luca Pacioli. After Pacioli’s work, this problem again appeared in the 1612 book Collection of the Pleasant and Entertaining Problems by the French mathematician Claude Gaspar Bachet Claude Gaspar de Meziriac. Bachet de Meziriac Claude Gaspar Bachet was born in BourgenBresse in Sa (15811638) voy (France). This was a region, which in different periods belonged to France, Spain or Italy. Therefore, the life of Claude Gaspar Bachet is connected with France, Spain and Italy of that period. He was a writer of books on mathemati cal puzzles and tricks that became a basis for all the books on mathematical puz zles. Bachet’s 1612 book Collection of the Pleasant and Entertaining Problems contains different numerical problems. Fibonacci’s “weighing” problem is one of them. This problem was formulated as follows: “What is the least number of weights used to weigh any whole number of pounds from 1 to 40 inclusively by using a balance, if the weights can be placed on both balance cups?” In the Russian historicalmathematical literature [169] Fibonacci’s “weighing” problem is known also un der the name of the BachetMendeleev problem in honor of Bachet de Meziriac and the famous Russian chemist Dmitri Mendeleev. However, this raises the question: why did the great Russian chemist become interested in the “weighing” problem? The answer to this question re quires some littleknown facts from the life of the great scientist. In 1892, Mendeleev was appointed the director Dmitri Mendeleev of the Russian Depot of Standard Weights, which, accord (18341907) Chapter 7 Algorithmic Measurement Theory 379 ing to Mendeleev’s initiative, was transformed in 1893 to the Main Board of Weights and Measures of Russia. Mendeleev remained its director until the end of his life. Thus, the final stage of Mendeleev’s life (since 1892 until his death in 1907) was connected with the development of metrology. During this period, Mendeleev was actively involved in various problems connected with measures, measurement and metrology; the “weighing” problem was one of them. Men deleev’s contribution to the development of metrology in Russia was so great, that he was named “the father of Russian metrology” and the “weighing” prob lem was named the BachetMendeleev problem. The essence of the problem consists of the following [169]. Suppose, we need to weigh on a balance any integervalued weight Q in the range from 0 up to Qmax by using the n standard weights {q1, q2, …, qn}, where q1=1 is a measurement unit; qi=ki×q1; ki is any natural number. It is clear that the maximal weight Qmax is equal to the sum of all the standard weights, that is, Qmax = q1+ q2+ …+ qn = (k1+ k2+…+ kn) q1. (7.9) We have to find the Optimal System of Standard Weights, that is, such standard weight system, which gives the maximal value of the sum (7.9) with a given measurement unit q1=1 among all possible variants. There are two variants of the solution to the BachetMendeleev problem. In the first case, the standard weights can be placed only on the free cup of the balance; in the second case, they can be placed on both cups of the balance. The Binary Measurement Algorithm with the binary system of standard weights {1, 2, 4, 8,…, 2n1} is the optimal solution for the first case. It is clear that the binary measurement algorithm generates the binary method of number representation: n −1 Q = ∑ ai 2 i , i =0 (7.10) where ai∈{0,1} is a binary numeral. Note that the binary numerals 0 and 1 have a precise measurement interpre tation. The binary numerals 0 and 1 encodes one of the two possible positions of the balance; thus, ai=1, if the balance remains in the initial position after putting a new standard weight on the free cup, otherwise ai=0. The ternary system of the standard weights {1, 3, 9, 27,…, 3n1} is the optimal solution to the second variant of the “weighing” problem when we can put the standard weights on both cups of the balance. It is clear that the ternary system of the standard weights results in the ternary method of number representation: n −1 Q = ∑ bi 3i , i =0 where bi is a ternary numeral that takes the values {1, 0, 1}. (7.11) Alexey Stakhov THE MATHEMATICS OF HARMONY 380 Hence, by solving the “weighing” problem, Fibonacci and his followers found a deep connection between measurement algorithms and positional methods of num ber representation. This is the key idea of algorithmic measurement theory [20, 21], which goes back to the “weighing” problem for its origin. Thus, from the measure ment problem we come unexpectedly to the positional method of number representa tion the greatest mathematical discovery of Babylonian mathematics. 7.4.3. Asymmetry Principle of Measurement The principles of Finiteness and Potential Feasibility of measurement, under lying the constructive measurement theory, are “external” with respect to the measurement. They have such a general character that there is a danger of their reduction to some trivial result (for example, to the above binary measurement algorithm that leads the measurement to the consecutive comparison of the measur able weight with the binary standard weights 2n1, 2n2, …, 20). For obtaining nontrivial results, the methodological basis of the con structive (algorithmic) theory of measurement should be added to by a cer tain general principle, which follows from the essence of measurement. Such principle follows directly from the analysis of the binary measurement algo rithm, which is the optimal solution to BachetMendeleev’s problem. + a) ... n1 2 Q n1 2 <Q 2 n1 2 n-2 ... 2 0 b) ... n1 n1 2 2 tQ 2 n2 ... 2 0 2n1 + ... c) n2 2 n1 2 Figure 7.3. Asymmetry Principle of Measurement 2 n2 ... 2 0 Chapter 7 Algorithmic Measurement Theory 381 We will analyze the above binary measurement algorithm by means of the use of the balance model (Fig. 7.3). This analysis allows for finding a measurement property of general character for any thinkable measurement, based on the com parison of the measurable weight Q with the standard weights. We shall examine the weighing process of the weight Q on the balance, by using the binary standard weights. On the first step of the binary algorithm the largest standard weight 2n1 is placed on the free cup of the balance (Fig. 6.3a). The balance compares the unknown weight Q with the largest standard weight 2n1. After the comparison, we can get two situations: 2n1<Q (Fig. 7.3a) and 2n1≥Q (Fig. 6.3b). In the first case (Fig. 7.3a), the second step is to place the next standard weight 2n2 on the free cup of the balance. In the second case (Fig.7.3b), the weigher should perform two operations. First, one should remove the previous standard weight 2n1 from the free cup of the balance (Fig. 7.3b), after that, the balance should return to the initial position (Fig. 7.3c). After returning the balance to the initial position, the next standard weight 2n2 should be placed on the free cup of the balance (Fig. 7.3c). As we can see, both cases differ in their complexity. In the first case, the weigher fulfils only one operation, that is, he adds the next standard weight 2n2 on the free cup of the balance. In the second case, the weigher’s actions are determined by two factors. First, he has to remove the previous standard weight 2n1 from the free cup of the balance, and after that, the balance is back to the initial position. After that, the weigher has to place the next standard weight on the free cup of the bal ance. Thus, for the first case the weigher has to perform only one operation. How ever, for the second case the weigher has to perform two sequential operations: (1) remove the previous standard weight 2n1 from the free cup; (2) place a new standard weight 2n2 on the free cup of the balance. In the second case, the weigher’s actions are more complicated in comparison with the first case because the weigher has to remove the previous standard weight taking into consideration the time necessary for the balance to return to its initial position. By synthesizing the optimal measurement algorithms, we have to consider these new data, which influence the process of measurement. We name this discov ered property of measurement, the Asymmetry Principle of Measurement [20]. 7.4.4. A New Formulation of the BachetMendeleev Problem Now, let us introduce the above measurement property into the Ba chetMendeleev problem. For this purpose we will examine a measurement as a process, running during discrete periods of time; let the operation to “add the standard weight” be carried out within one unit of discrete time Alexey Stakhov THE MATHEMATICS OF HARMONY 382 and the operation to “remove the standard weight” (which is followed by re turning the balance into the initial position) be carried out within p units of discrete time, with p∈{0,1,2,3,…}. It is clear that the parameter p simulates the Inertness of the balance. Note that the case p=0 corresponds to the “ideal situation” when we neglect the inertness of the balance. This case corresponds to the classical BachetMendeleev problem. For the other cases p >0, we have some new variants of the BachetMendeleev problem, which are studied in the algorithmic measurement theory [20, 21]. Below we study other generalizations of the BachetMendeleev problem. 7.5. Mathematical Model of Measurement 7.5.1. A Notion of the “Indicator Element” If we put forward a problem to create the Algorithmic Measurement The ory, we should define more exactly, what is a Measurement, what is the Purpose of Measurement, what is a Measurement Algorithm, and How are the Measure ment Algorithms carried out. First of all, we should note that, if we wish to measure something, we should know the range of the measurable magnitude. When we discuss “mathematical measurement,” we are distracted from the physical nature of measurable magni tude. We will present a Range of Measurable Magnitude in the form of the geo metric line segment AB. Clearly the measurable magnitude is one of the possible sizes of the magnitude that belong to the range of measurement, that is, prior to the measurement, a certain “indeterminacy” about the measurable size exists be cause otherwise the measurement would be senseless. We will represent this situ ation of “indeterminacy” by the “unknown” point X, which belongs to the line segment AB. Now, let us formulate a purpose for the measurement. The purpose of the measurement is to find the length of the “unknown” segment AX. In practice this purpose is carried out by means of special devices, for example, “balances” or “comparators.” The “comparators” perform a comparison of the measurable magnitude with some “measures,” generated from the “measure unit.” In de pendence on the result of comparison, the device “indicates” the result of com parison in the form of a binary signal 0 or 1. Thus, the essence of the measure ment consists of Consecutive Comparisons of the measurable magnitude with some “measures,” which are formed at each step of the measurement. More Chapter 7 Algorithmic Measurement Theory 383 over, each step of the measurement depends on the results of comparison on the preceding steps. In order to model the process of comparison of the measurable segment AX with the “measures,” we introduce the important concept of the Indicator Element (IE), which is an original model of the “comparator” or “balance,” the basic tool of any measurement. Suppose that every IE can be put to any “known” point C of the segment AB. The indicator element provides the information about the mutual positioning the “unknown” point X and the “known” point C. If the IE is to the right of the point X, it “generates” the binary signal 0; otherwise, the binary signal 1. 7.5.2. The (n, k, S)algorithms Some Conditions or Restrictions S (that follow, for example, from the “Princi ple of Asymmetry of Measurement”) can be imposed on the measurement pro cess; thus, by means of the restrictions S the “inertness” property of the balance can be taken into consideration. A C C C X C C B By using a concept of the in ... ... dicator elements, we can describe Figure 7.4. A geometric model of measurement the process of measurement in terms of geometry. Suppose we have a line segment AB with some “unknown” point X (Fig. 7.4). The problem is to find the length of line segment AX. This is performed by means of the k “Indicator Elements” (IE). After we enclose the jth IE (j=1,2,3,…,k) at some point Cj, the line segments AX and ACj are compared, i.e. the relations “less than” (AX< ACj) or “greater than or equal” (AX≥ACj) are defined. The relations “less than” and “greater than or equal” are encoded by the binary numerals 0 and 1, which are generated by the IE in the process of measurement. The problem of the measurement of a line segment AX by the k IE’s is reduced to decreasing the Inde terminacy Interval about X according to the IE“indications.” A measurement procedure consists of the Measurement Steps: in the first step the k IE’s are enclosed to some points of the line segment AB, and the indeterminacy interval about X is diminished to the line segment A1B1 ac cording to the IE“indications”; in the next step the IE’s are enclosed to the points of the line segment A1B1, and so on. The restrictions S are imposed on the measurement procedure. For the given restrictions S and the given n and k, the system of formal rules, which strictly define the points of the k IE’s enclosures for each step in dependence on the IE“indications” on the preced ing steps, is called the (n,k,S)algorithm. 1 2 j j+1 k Alexey Stakhov THE MATHEMATICS OF HARMONY 384 7.5.3. The Optimal (n, k, S)algorithms A model of the indicator elements reduces the measurement problem to the problem of searching the point X (its coordinate is equal to the length of the line segment AX) within the line segment AB by means of the use of the k IE’s in n steps. The above approach allows one to use the methods of onedimensional search for the synthesis of the optimal measurement algorithms [20]. Let us consider the (n,k,S)algorithm, which acts on the line segment AB. Some “indeterminacy interval” ∆i, which contains the point X on the last step of the algorithm, is the result of the (n,k,S)algorithm for the certain point X. Let us consider the action of the algorithm for all possible points X∈AB. As a result we obtain the set of line segments {∆i} called the Partition P of the Line Segment AB, i.e. P={∆1, ∆2 , …, ∆i , …, ∆N }. (7.12) For the general case, the partition (7.12) consists of the N line segments of the kind ∆i. They are not equal to one another in the general case. Besides, there is the following correlation for the partition P: AB=∆1+∆2+…+∆N. (7.13) We choose from the partition (7.13) the largest line segment ∆max and define the effectiveness of the (n,k,S)algorithm by means of the line segment ∆ max . Consider the ratio AB T= , (7.14) ∆ max which is called the (n,k)exactness of the (n,k,S)algorithm. In accordance with (7.14), each (n,k,S)algorithm is characterized by the number T, which is some numerical evaluation of the effectiveness of the (n,k,S) algorithm. The availability of such a number T allows one to compare the dif ferent (n,k,S)algorithms according to their effectiveness. By using the defini tion (7.14), we can introduce the notion of an Optimal (n,k,S)algorithm. Definition 7.1. For some given n, k, and S, the (n,k,S)algorithm is called optimal, if it provides a maximal value of the (n,k)exactness T given by (7.14) among all possible variants. Note that this definition is based on the socalled Minimax Principle [20]. This principle is used widely in the modern theory of optimal systems. Ac cording to this principle we refer some strategy to the “optimal” strategy, if it provides the maximal value of the efficiency criterion for the worst case. It is easy to prove [20] that the optimal (n,k,S)algorithm divides the line segment AB into N equal parts, i.e. Chapter 7 Algorithmic Measurement Theory 385 ∆1 = ∆ 2 = ... = ∆ N = ∆. (7.15) For the case (7.15) we have the following expression for the (n,k)exactness: AB T= = N. (7.16) ∆ Thus, in accordance with (7.16) the ( n, k ) exactness of the optimal (n,k,S) algorithm coincides with the number N of the quantization levels that are pro vided by the optimal (n,k,S)algorithm. 7.6. Classical Measurement Algorithms Over the millennia, the practice of human measurement found a number of measurement algorithms that are widely used today in mathematics and mea surement technology. The most common among them include the following: Binary Algorithm, Counting Algorithm, Ruler Algorithm. We will study these algorithms by using the above “indicator” model of measurement (Fig. 7.4). 7.6.1. The “Binary” Algorithm The essence of the “Binary” Algorithm is demonstrated in Fig. 7.5. It con sists of 3 steps (n steps in the general case) and uses only IE (k=1). By using 3 steps, the algorithm divides the initial line segment [0,8] into 8 equal parts. The First Step is to enclose the IE to the middle of the initial line segment [0,8], i.e. to point 4. After the first step there appear two situations in depen dence on the IE“indication,” [0,4] and [4,8]. The Second Step: (a) If the IE“indication” in point 4 is equal to 0 (the IE appears to the left), this means that the “unknown” point X is at the line segment [0,4]. For this situation the second step of the “binary” algorithm is to enclose the IE to the middle of the line segment [0,4], that is, to point 2. (b) If the IE“indication” in point 4 is equal to 1 (the IE appears to the right), this means that the “unknown” point X is at the line segment [4,8]. For this situation the second step of the “binary” algorithm is to enclose the IE to the middle of the line segment [4,8], i.e. to point 6. It is clear that after the second step four situations may appear as shown in Fig. 7.5. Alexey Stakhov THE MATHEMATICS OF HARMONY 386 8 4 0 1st step 0 2 4 4 6 8 2d step 0 1 2 2 3 4 4 5 6 3d step 6 7 8 Figure 7.5. The “binary” algorithm The Third Step. We can see from Fig. 7.5 that after the second step of the “binary” algorithm we have the four “indeterminacy intervals” [0,2], [2,4], [4,6], [6,8]. The third step of the “binary” algorithm is to enclose the IE at the middle of one of the “indeterminacy intervals” [0,2], [2,4], [4,6], [6,8]. It follows from Fig. 7.5 that the “binary” algorithm provides the following partition of the initial line segments [0,8]: P={[0,1], [1,2], [2,3], [3,4], [4,5], [5,6], [6,7], [7,8]}, that is, it divides the line segment [0,8] into 8 equal parts. It is clear that the above “binary” algorithm is the (3,1,S)algorithm and its (3,1)exactness is equal to T=23=8. It is easy to prove that the nstep “binary” algorithm provides the (n,1)exactness given by the following formula: T = 2n . (7.17) 7.6.2. The “Counting” Algorithm Now, let us consider the socalled Counting Algorithm widely used in mea surement practice. The “counting” algorithm (see Fig. 7.6) consists of 3 steps (n steps in the general case) and uses only the IE (k=1). It divides the initial line segment [0,4] into 4 equal parts in 3 steps. The First Step is to enclose the IE at the point 1. After the first step there appear two situations, [0,1] and [1,4], depend on the IE“indication” in point 1. Chapter 7 Algorithmic Measurement Theory 387 The Second Step: (a) If the IE“indication” in 0 1 point 1 is equal to 0 (the IE appears 4 to the left), then the “unknown” 1 2 2d step point X is at the line segment [0,1]. 1 2 For this situation the measurement process is over because the Xcoor 3 4 2 3d step dinate (X∈[0,1]) is defined with “exactness” equal to measurement Figure 7.6. The “counting” algorithm unit 1. (b) If the IE“indication” in point 1 is equal to 1 (the IE appears to the right), this means that the “unknown” point X is at the line segment [1,4]. For this situation the measurement process will continue and the IE is en closed at point 2. The Third Step. We can see from Fig. 7.6 that on the third step the IE is enclosed at point 3, which is from point 2 on the distance equal to the measure ment unit 1. From the example in Fig. 7.6 we can find the following general rule for the “counting” algorithm. If on the ith step of the “counting” algorithm, which acts at the line segment AB, the IE was enclosed at point C, then we have two situations on the (i+1)th step. If the IE appears at point C to the left, then the measurement procedure is over. If the IE appears at point C to the right, then we get the “indeterminacy interval” CB and the IE on the (i+1)th step is en closed at point C′=C+1. It is clear that the above “counting” algorithm is the (3,1,S)algorithm and its (3,1)exactness is equal to T=4. It is easy to prove that the nstep “counting” algorithm has the (n,1)exactness given by the following formula: 1st step 0 1 T = n +1. 4 (7.18) 7.6.3. The “Ruler” Algorithm Let’s consider one more example of the measurement algorithm widely used in measurement practice. The latter underlies the traditional measure ment ruler. We call this algorithm the “Ruler” Algorithm. The essence of the latter is demonstrated in Fig. 7.7. The “ruler” algorithm is performed in 0 1 3 2 4 1 step and uses the k IE (the 3 IE in Fig. 7.7). The algorithm divides the initial Figure 7.7. The “ruler” algorithm line segment into Alexey Stakhov THE MATHEMATICS OF HARMONY 388 (7.19) T = k +1 equal parts (into 4 parts in Fig. 7.7). The formula (7.19) gives the (1, k)exactness of the “ruler” algorithm. We can see that the latter refers to a class of the (1,k,S)algorithms. 7.6.4. The Restriction S The above definition of the (n,k,S)algorithm includes a concept of the Restriction S. Clearly, any “restriction” limits the efficiency of the algorithm. However, each “restriction” (as show below) can result in rather unusual measurement algorithms, which are of theoretical and practical interest. We start with the elementary “restrictions,” which can be found in the comparative analysis of the classical measurement algorithms, in particular, the “binary” algorithm and the “counting” algorithm. Compare two 3step (n,k,S)algorithms, shown in Figs. 7.5 and 7.6. Both algorithms consist of 3 steps and use only the IE (k=1). What is the difference between them? The difference consists in the character of the movement of the IE along the line segment AB. Consider the restriction S for the “binary” algorithm in Fig. 6.6. Remember that the restriction S is imposed upon the movement of the IE along the line segment AB. We can see that the IE for the “binary” algorithms may be enclosed at each step to the left or to the right of the “unknown” point X. One may say that the movement of the IE for this case is performed with out any “restrictions.” Denote the restriction S for the “binary” algorithm by S≡0. We will call the class of (n,k,S)algorithms, which each individually satisfy the restriction S≡0, the (n,k,0)algorithms. Consider the restriction S for the “counting” algorithm. We can see that for this case the IE is moving along the line segment [0,4] in only one direc tion, namely, from the point 0 to the point 4. We denote the restriction of this kind by S≡1 and call the class of (n,k,S)algorithms satisfying the restriction S≡1, the (n,k,1)algorithms. Note that the restrictions S≡0 and S≡1 are not the only possible “re strictions.” Below we will study the “restriction” S, which follows from the above formulated Asymmetry Principle of Measurement. It is important to emphasize, that each “restriction” results in the development of one or an other class of the new measurement algorithms, which can be of theoretical or practical interest; therefore, a search of reasonable “restrictions” impos able on the measurement algorithm, is an important problem in algorithmic measurement theory. Chapter 7 Algorithmic Measurement Theory 389 7.6.5. Connection between Measurement Algorithms and Positional Number Systems There is a deep connection between measurement algorithms and positional methods of number representation. In fact, the “binary” algorithm generates the binary representation of numbers: (7.20) A = an −1 2n −1 + an −2 2n −2 + ... + ai 2i + ... + a1 21 + a0 20 , where ai is a binary numeral {0,1}; 2i is the weight of the ith digit (i=0,1,2, …, n1). The formula (7.20) has the following “measurement” interpretation. The binary numerals an1, an1, …, a0 are the IE“indications” on the first, second, ... , nth step of the “binary” algorithm, respectively. One can readily see that the “counting” algorithm “generates” the following method of number representation that underlies the Euclidean definition of natural number: N = 1 + 1 + ... + 1. N (7.21) The latter is well known as a Unitary Code of the number N. 7.7. The Optimal Measurement Algorithms Originating Classical Position al Number Systems 7.7.1. A General Method for the Synthesis of the Optimal (n,k,S) algorithms For the synthesis of the optimal (n,k,S)algorithm we use the Method of Recurrence Relations [170]. To obtain the recursive relation for (n,k)exact ness T of the optimal (n,k,S)algorithm, we first formulate an Inductive As sumption: for any arbitrary n, k and S there is the optimal (n,k,S)algorithm, which divides the line segment AB into T equal line segments of length D, where T is the (n,k)exactness of the optimal (n,k,S)algorithm. For a given restriction S, the (n,k)exactness T depends on n and k. We can present this dependence in the form of some function of n and k, i.e. T=F(n,k). It is clear if the length D=1, then the line segment (7.22) AB=T=F(n,k). (7.23) Alexey Stakhov THE MATHEMATICS OF HARMONY 390 Consider the first step of the optimal (n,k,S)algorithm, which acts on the line segment AB (Fig. 7.8). A C1 C2 Cj ... Cj+1 Ck B ... Figure 7.8. The first step of the optimal (n, k, S)algorithm Let the first step be to enclose the k IE’s at some points C1,C2,…,Cj,Cj+1, …, Ck. Then, according to the IE“indications” the following (k+1) different situations may result: (1) X ∈ AC1 ; ( 2 ) X ∈C1C2 ; ...; ( j + 1) X ∈C jC j +1 ; ...; ( k + 1) X ∈Ck B. (7.24) The second step of the optimal (n,k,S)algorithm is to enclose the IE’s at the points of the new “indeterminacy interval,” which is one of the line segments (7.24). It is clear that for this situation we can use the optimal (n1)algorithm. However, the number of the IE’s on the second step depends on the restriction S. Suppose, that in accordance with the restriction S we can use only the r IE’s (r≤k) on the second step of the algorithm. Note that the number r may vary for the different situations (7.24). Suppose that the line segment CjCj+1 is the “indeterminacy interval” on the second step. Since on the second step we have r IE’s and (n1) steps of the algo rithm, we can use the optimal (n1,r,S)algorithm on the line segment CjCj+1. In accordance with the inductive assumption, the (n1,r,S)algorithm will divide the line segment CjCj+1 into F(n1,r) equal line segments of length ∆=1, i.e. CjCj+1 = F(n1,r). Using the equality (7.13), we can write: (7.25) AB=AC1+ C1C2+…+ CjCj+1+…+CkB. (7.26) Using (7.23) and (7.25), we can get the recursive relation for the calcula tion of the function (7.23). 7.7.2. The Optimal (n,k,0)algorithms Now, we can use the above method for the synthesis of the optimal (n,k,0) algorithm. Recall that the restriction S≡0 means what each IE may be en closed at any arbitrary point of the line segment of AB on each step. According to the above method, one of the situations (7.24) appears after the first step of the optimal (n,k,0)algorithm. Consider the situation: X∈CjCj+1, which can appear after the first step of the algorithm. Then in accordance with the restric tion S≡0, the second step of the optimal (n,k,0)algorithm is to enclose all k IE’s at the points of the new line segment CjCj+1. If we use the (n1,k,0)algorithm on the Chapter 7 Algorithmic Measurement Theory 391 “indeterminacy interval” CjCj+1, then in accordance with the inductive assumption we can divide the line segment CjCj+1 into F(n1,k) equal parts ∆=1, i.e. (7.27) CjCj+1 = F(n1,k). The expression (7.27) is valid for any arbitrary line segment from the set (7.24). Then, by using (7.26), we obtain the following recursive relation for the calculation of the function F(n,k): F(n,k)=(k+1)F(n1,k). (7.28) The recursive function (7.28) can be expressed in explicit form. For this purpose, we can use the recursive relation F(n1,k)=(k+1)F(n2,k) for the representation of (7.28) in the following form: F(n,k)=(k+1)(k+1)F(n2,k). By continuing this process, we can represent the function (7.28) in the fol lowing form: F(n,k)=(k+1)n1F(1,k). (7.29) Thus, according to (7.29) the solution to the problem is reduced to obtain ing the expression for the function F(1,k). The latter is the expression for the (1,k)exactness of the optimal (1,k,0)algorithm. Remember that the (1,k,0) algorithm consists of 1 step. This step is to enclose the k IE’s at the points of the line segment AB. It is easy to prove that in this case the optimal solution is to divide the line segment AB into (k+1) equal parts. It follows from this consid eration that the (1,k)exactness of the optimal (1,k,0)algorithm is equal to F(1,k)=k+1. (7.30) By substituting (7.30) into (7.29), we obtain the expression for the function (7.28) in the explicit form: (7.31) F(n,k)=(k+1)n . The expression (7.31) allows one to formulate the essence of the optimal (n,k,0) algorithm. Each step of the algorithm is to divide the initial “indeterminacy inter val” (the line segment AB) and all the next “indeterminacy intervals” into (k+1) equal parts. As a result of the application of the algorithm, we obtain the formula (7.31), which characterizes the effectiveness of the optimal (n,k,0)algorithm. 7.7.3. Special Cases of the Optimal (n,k,0)algorithm Consider some special cases of the optimal (n,k,0)algorithm. For the case k=1 the formula (7.31) is reduced to the formula (7.17), which expresses Alexey Stakhov THE MATHEMATICS OF HARMONY 392 the (n,1)exactness of the “binary” algorithm. For the case n=1 the formula (7.31) is reduced to the formula (7.19), which expresses the (1,k)exactness of the “rul er” algorithm. It follows from this examination that the wellknown measure ment algorithms, “binary” algorithm and “ruler” algorithm, are partial cases of the optimal (n,k,0)algorithm. Similar to the “binary” algorithm, which generates the “binary” number system (7.20), the optimal (n,k,0)algorithms generate the positional number systems with radix R=k+1, i.e. (7.32) A=an1Rn1+ an2Rn2+…+aiRi+…+ a1R1+ a0R0, where ai∈{0,1,2,3,…,k} is a numeral of the ith digit. The expression (7.32) has the following “measurement” interpretation. The numeral an1 in (7.32) encodes the “indications” of the IE’s on the first step of the optimal (n,k,0)algorithm according to the rule: 0 = 000...00 1 = 100...00 2 = 110...00 ... (7.33) k − 1 = 111...10 k = 111...11 The numeral an2 encodes the “indications” of the IE’s on the second step of the algorithm and the numerals a1 and a0 encode the “indications” of the k IE’s according to the rule (7.33) on the (n1)th and nth steps of the optimal (n,k,0)algorithm, respectively. It is clear that for k=9 the optimal (n,9,0)algorithm generates the deci mal system and for k=59 the Babylonian sexagesimal system. It follows from this consideration that the optimal (n,k,0)algorithms generate all historical ly wellknown positional systems. 7.8. Optimal Measurement Algorithms Based on the Arithmetical Square 7.8.1. The Optimal (n,k,1)algorithms Consider the class of the (n,k,1)algorithms. Remember that the restriction S≡1 means that the indicator elements (IE) are moving along the line segment AB only in the direction from point A to point B. This means, if any IE on any step is found to the right of the “unknown” point X, then this IE will “leave the field,” Chapter 7 Algorithmic Measurement Theory 393 i.e. this IE cannot be used for further measurement. Note that the above “count ing” algorithm (Fig. 7.6) is the simplest example of the restriction S≡1. Let the first step of the optimal ( n, k,1) algorithm be to enclose the k IE’s to the certain points C1,C2,…,Cj,Cj+1, …, Ck of the line segment AB (Fig. 7.8). Consider the situations (7.24) that can appear after the first step of the algo rithm. Consider the situation X∈C jCj+1. (7.34) It is clear for the situation (7.34) that the j IE’s are on the left of the “un known” point X and the rest, the (kj)th IE’s, are on the right of point X. In accordance with the restriction S≡1, the (k j) IE’s, which are on the right of the “unknown” point X, cannot be used for further measuring. This means that the second step of the algorithm is to enclose to the line segment CjCj+1 only those j IE’s which are found on the left of the “unknown” point X after the first step of the algorithm. That is why we can only use the optimal (n 1,j,1)algorithm for the situation (7.34). In accordance with the inductive as sumption, the optimal (n1,j,1)algorithm can divide the line segment CjCj+1 into F(n1,j) equal parts ∆=1, that is, CjCj+1=F(n1,j). Consider the situations (7.35) X∈AC1 and (7.36) (7.37) X∈CkB. For the situation (7.36) all IE’s are to the right of the point X after the first step. This means that all IE’s “leave the field” and the process of the measure ment is over. It is clear from this examination that the line segment AC1 must be the line segment of a single length, i.e. AC1=1. (7.38) For the situation (7.37) all IE’s are to the left of point X after the first step of the algorithm and they can be enclosed to the points of the line segment CkB on the second step of the algorithm. Using the optimal (n1,k,1)algorithm, we can divide the line segment CkB into F(n1,k) equal parts ∆=1, i.e. CkB=F(n1,k). (7.39) Using the expressions (7.26), (7.35), (7.38), (7.39), we obtain the follow ing recursive relation for the calculation of the function F(n,k), which gives the (n,k)exactness of the optimal (n,k,1)algorithm: F(n,k)=1+F(n1,1)+F(n1,1)+F(n1,2)+…+F(n1,j)+…+F(n1,k1)+F(n1,k). (7.40) Alexey Stakhov THE MATHEMATICS OF HARMONY 394 Using the recursive relation (7.40), we can write the following recursive for mula for the calculation of the function F(n, k1): F(n,k1)=1+F(n1,1)+F(n1,1)+F(n1,2)+…+F(n1,j)+…+F(n1,k1). (7.41) Comparing the expressions (7.40) and (7.41), we can rewrite the recursive relation (7.40) in the following form: F(n,k)= F(n,k1)+ F(n1,k1). (7.42) Calculating the values of the function F(n, k) for the special cases n=0 and k=0, we clearly have: F(0,k)=1 (7.43) F(n,0)=1. (7.44) For an explanation of the “physical sense” of the expressions (7.43) and (7.44) we can examine the optimal (n, k,1)algorithm for cases n=0 and k=0. The case n=0 means that we do not have any step for the measurement and, hence, the “indeterminacy interval” (the line segment AB) cannot be decreased. Thus, the expression (7.43) is valid. Now, suppose that k=0. This means that we do not have any “Indicator Elements” for the measurement and the “indeterminacy interval” (the line seg ment AB) cannot be decreased. Thus, the expression (7.44) is valid. 7.8.2. Arithmetical Square Using the recursive relation (7.42) with the initial terms (7.43) and (7.44), we can construct a table of the values of the function F(n,k) (Table 7.1). We can see that Table 7.1 coincides with the wellknown Arithmetical Square or Pascal Triangle and that the terms F ( n, k ) are Binomial Coefficients, that is, F ( n, k ) = C nk+ k = C nn+ k . (7.45) Table 7.1. Arithmetical Square k/n 0 1 0 1 1 1 1 2 2 1 3 3 1 4 4 1 5 5 1 6 ... ... ... 2 1 3 6 10 15 21 ... n 1 F ( n, 1) F ( n, 2 ) F ( n, 3 ) 3 1 4 10 20 35 56 ... 4 1 5 15 35 70 126 ... 5 1 6 21 56 126 252 ... F ( n, 5 ) # k # 1 # F (1, k ) # F ( 2, k ) # F ( 3, k ) # F ( 4, k ) # F ( 5, k ) ### ... # F ( n, k ) F ( n, 4 ) Chapter 7 Algorithmic Measurement Theory 395 7.8.3. The Optimal Measurement Algorithms Based on Arithmetical Square The arithmetical square (Table 7.1) allows to demonstrate the optimal (n,k,1)algorithm. In fact, for the given n and k, the value of the function F(n,k) is at the intersection of the nth column and kth row of the arithmetical square. The coordinates of the k IE’s, which should be enclosed to the points C1,C2,…,Cj,Cj+1, …, Ck of the initial line segment AB, are in the nth column of the arithmetical square, that is, AC1 = 1, AC2 = F ( n,1) , ..., AC j = F ( n, j − 1) , AC j +1 (7.46) = F ( n, j ) , ..., AC k = F ( n, k − 1) . If after the first step of the algorithm the j IE’s are to the left of point X and the rest (kj) IE’s are to the right of X, the “indeterminacy interval” about X is dimin ished up to the line segment CjCj+1. The length of the latter is equal to the binomi al coefficient F(n,j). This binomial coefficient is at the intersection of the (n1)th column and the jth row of the arithmetical square. To obtain the binomial coeffi cient F(n,j) we have to move from the initial coefficient F(n,k) by one column to the left and by (kj) rows upwards. On the second step we have to accept the point Cj as the new beginning of coordinates. Here the second step is to enclose j IE’s to the points D1, D2, …, Dj of the new “indeterminacy interval” CjCj+1. The coordinates of the points D1, D2, …, Dj with respect to the point Cj the new begin ning of coordinates are in the (n1)th column of the arithmetical square, that is, C j D1 = 1, C j D2 = F ( n − 1,1) , C j D3 = F ( n − 1, 2) , ...,C j D j = F ( n − 1, j − 1) . (7.47) This process continues up to the exhaustion either of all measurement steps or all IE’s. 7.8.4. The Example of the Optimal (n, k, 1)algorithm Consider the optimal (3, 3, 1)algorithm (Fig. 7.9). The algorithm con sists of 3 steps (n steps in the general case) and uses 3 IE’s (k IE’s in the general case). The value of the initial “indeterminacy interval” is the binomi al coefficient F(3, 3)=20, which is at the intersection of the 3d column and 3 d row of the arithmetical square. The first step of the algorithm at line segment [0,20] is to enclose 3 IE’s to the points 1, 4, 10 (Fig. 7.9). In accordance with the IE“indications,” four situations may appear after the first step: [0,1], [1,4], [4,10], [10,20]. The second step: (1a) For the situation [0,1] the process of measurement is over because all IE’s are on the right of the point X. Alexey Stakhov THE MATHEMATICS OF HARMONY 396 0 1 4 10 20 (1b) For the situa tion [1,4] we have the 0 1 only 1 IE, which can be (1a) 1 2 4 enclosed at point 2. (1b) (1c) For the situa 7 10 4 5 tion [4,10] we have 2 (1c) 16 10 11 13 20 IE’s, which can be en (1d) closed to points 5 and 7. Figure 7.9. The optimal (3, 3, 1)algorithm (1d) For the situa tion [10,20] we have 3 IE’s. For this case the coordinates of three points of the IE’s enclosing line segment [10,20] with respect to the point 10, the new begin ning of coordinates, are in the second column of the arithmetical square (Table 7.1) above binomial coefficient 10. Those are the binomial coefficients 1, 3, 6. By summing these numbers with the number 10, we obtain the coordinates of the points of the IE’s enclosed on the second step, namely: 11=10+1, 13=10+3, and 16=10+6. The third step. After the second step, the following situations will appear: [1,2], [2,4], [4,5], [5,7], [7,10], [10,11], [11,13], [13,16], [16,20]. Note that for the situa tions [1,2], [4,5], [10,11] the measurement process is over on the second step. It is clear from Fig. 7.9 that the optimal (3,3,1)algorithm can divide the initial line segment [0,20] into 20 equal parts. 7.9. Fibonacci Measurement Algorithms 7.9.1. The Optimal Measurement Algorithms Based on the Fibonacci pnumbers Return to the “indicator” measurement model (Fig. 7.4) and try to intro duce the above Asymmetry Principle of Measurement into this model. The following “restriction” to the movement of the IE along the line segment AB follows from the Asymmetry Principle of Measurement. Suppose that the IE is enclosed at the point C on the first step of the n step measurement algorithm (Fig. 7.10). Thereafter, two situations (a) and (b) appear after the first step, as shown in Fig. 7.10. It is clear that for the situation (a) we can enclose the IE at any point of the “indeterminacy interval” CB on the next step. For the situation (b) Chapter 7 Algorithmic Measurement Theory 397 B C A we do not have the right to en close the IE to the points of the “indeterminacy interval” AC, be B C cause on the first step the balance (a) F(n1) moves to the opposite position A C and we need p units of discrete (b) F(np1) time for returning the balance Figure 7.10. The first step of the Fibonacci into the initial position. Thus, measurement algorithm the restriction S, which follows from the Asymmetry Principle of Measurement, consists of the fact that for the situation (b) it is forbidden to enclose the IE to the points of the line segments AC during the next p steps of the algorithm. We can use the above restriction for synthesis of the optimal measurement algorithm. Here we introduce the following “inductive assumption.” Suppose that for the given n≥0 and p≥0 there is the optimal nstep measurement algorithm, which divides the initial line segment AB into Fp(n) equal parts of ∆=1, that is, (7.48) AB=Fp(n). Suppose that the first step of the algorithm is to enclose the IE at the point C (Fig. 7.10). Then, after the first step two situations can appear. Line segment CB is the “indeterminacy interval” for situation (a). According to the restriction S we can enclose the IE at any point of CB on the second step. By using the optimal (n1)step algorithm in this situation we can divide the line segment CB into the Fp(n1) equal parts ∆=1 and, hence, CB = Fp(n1). (7.49) Line segment AC is the “indeterminacy interval” for the situation (b). According to the restriction S, we cannot enclose the IE to the points of the line segment AC during the next p steps of the algorithm. Let us consider the situation (b) for two possible cases: (1) p≥n1 (2) p<n1. It is clear that all steps of the algorithm for case (1) will be “exhausted” much earlier before the “prohibition” to enclose the IE to the points of the line segment AC will be taken off. This means that for situation (1) the mea surement is over after the first step. That is why line segment AC should be the line segment of unit length, that is, AC=1. (7.50) Consider the situation (b) for case (2). For this case we have the right to enclose the IE to the points of the line segment AC only after p steps. Then, acting by the optimal (np1)algorithm we can divide the line segment AC into the Fp(np1) equal parts D=1, that is, Alexey Stakhov THE MATHEMATICS OF HARMONY 398 AC=Fp(np1). (7.51) Taking into consideration that AB=AC+CB, we can write the following expression for the function Fp(n) for cases (1) and (2), respectively: Fp(n)= Fp(n1)+1 with p≥n1 (7.52) Fp(n)= Fp(n1)+ Fp(np1) with p<n1. (7.53) Note that the recursive formula (7.53) coincides with the recursive for mula for the Fibonacci pnumbers introduced in Chapter 4. That is why the measurement algorithms that correspond to the mathematical formulas (7.52) and (7.53) were called Fibonacci Measurement Algorithms [20]. Re member that Fibonacci pnumbers is a numerical sequence that is given by the following recursive relation Fp(n)=Fp(n1)+ Fp(np1) at the seeds: (7.54) (7.55) Fp(1)= Fp(2)=…= Fp(p+1)=1. Consider in greater detail the mathematical formulas (7.52) and (7.53). For the case p≥n1 the Fibonacci measurement algorithms are described by the for mula (7.53). The recursive relation (7.52) may be expressed in the explicit form. In fact, by decomposing the function Fp(n1) in (7.52) and all the fol lowing functions Fp(n2), Fp(n3), Fp(1) in accordance with the same recur sive relation (7.52), we obtain the following expression: Fp(n)=n+Fp(0). (7.56) It follows from the “physical” reasoning that Fp(0)=1 and then the expres sion (7.56) may be represented in the following “explicit” form: Fp(n)=n+1. (7.57) Note that the formula (7.57) generates the natural numbers. By joining the formulas (7.57) and (7.53), we can write the following expression for the function Fp(n) with p ≥ n − 1 n = 1 Fp (n) = . (7.58) Fp ( n − 1) + Fp ( n − p − 1) with p < n − 1 Consider some special cases of the formula (7.58) for the different values of p. Let p=0. For this case the formula (7.58) is reduced to the following: F0(n)=2 F0(n1) (7.59) F0(0)=1. (7.60) It is clear that the recursive formula (7.59) at seed (7.60) generates the “binary” sequence: Chapter 7 Algorithmic Measurement Theory 399 1, 2. 4. 8, 16, …, F0(n)=2n1. It is clear that the Fibonacci measurement algorithm that corresponds to this case is reduced to the classical “binary” algorithm. Let p=∞. This means that the IE “leaves the field” when it is on the right of the point X. For this case the formula (7.58) takes the form of the expres sion (7.57) that generates natural numbers. It is clear that the classical “count ing” algorithm is the optimal Fibonacci measurement algorithm correspond ing to this case. Table 7.2 gives the values of the (n,1)exactness (7.58) of the optimal Fibonacci algorithms for the different values p. The expression (7.58) Table 7.2. The (n,1)exactness of the Fibonacci gives the (n,1)exactness measurement algorithms of the optimal nstep Fi bonacci algorithm for the Fp ( n ) / n 1 2 3 4 5 6 7 8 9 F0 ( n ) 2 4 8 16 32 64 128 256 512 restriction S that follows from the Asymmetry Prin F1 ( n ) 2 3 5 8 13 21 34 55 89 ciple of Measurement. We F2 ( n ) 2 3 4 6 9 13 19 28 41 find the expression for the F3 ( n ) 2 3 4 5 7 10 14 19 26 standard weights for Fi # # # # # # # # # # bonacci algorithms in the 8 9 10 F∞ ( n ) 2 3 4 5 6 7 form: {Wp(1), Wp(2), …, Wp(i), …, Wp(n)}, where Wp(i) is the ith standard weight. It is easily proven that the standard weight Wp(n) is given by the following expression: Wp(n)= Wp(n1)+Wp(np1) with n>p+1 (7.61) Wp(1)= Wp(2)= … = Wp(p+1)=1. (7.62) Table 7.3 gives different variants of the standard weight systems correspond ing to different values of p. A comparison of the expressions (7.61) and (7.62) with expressions (7.52) and (7.53) shows that Table 7.3. Standard weights for the Fibonacci the standard weights measurement algorithms Wp(n) for Fibonacci’s Wp ( n ) / n 1 2 3 4 5 6 7 8 9 measurement algo W0 ( n ) 1 2 4 8 16 32 64 128 256 rithms coincide with W1 ( n ) 1 1 2 3 5 8 13 21 34 the pFibonacci num W2 ( n ) 1 1 1 2 3 4 6 9 13 bers [20], in particular, W3 ( n ) 1 1 1 1 2 3 4 5 7 with the classical Fi # # # # # # # # # # bonacci numbers for W n 1 1 1 1 1 1 1 1 1 ( ) ∞ the case p = 1 . Alexey Stakhov THE MATHEMATICS OF HARMONY 400 7.9.2. The Example of the Fibonacci Measurement Algorithm Here we examine the optimal Fibonacci’s measurement algorithm given by the expression (7.58) for the case p=1. Let n=5. Consider the 5step Fibonacci algorithm (Fig. 7.11) corresponding to the case p=1. It follows from Table 7.2 that the above 5step Fibonacci algorithm provides the (5,1)exactness F1(5)=13. This means that the optimal 5step Fibonacci algorithm divides the line segment [0,13] into 13 equal parts. Thus, the given Fibonacci measurement algorithm consists of 5 steps and uses 5 standard weights {1, 1, 2, 3, 5} (see Table 7.3). Let us consider the first 3 steps of the given algorithm. The first step. We can use the first standard weight 5 for comparison with the measurable weight. This means that in the terms of the “indicator” model we have to enclose the IE at the point 5 of the line segment [0,13] (Fig. 7.11). We can see that the first step consists in the division of the line segment [0,13] in the Fibonacci ratio: 13=5+8. Two situations (a) and (b) (Fig. 7.11) can appear after the first step. 0 5 (a) (b) 13 5 0 8 13 5 forbidden 8 (c) 5 (d) 10 13 10 13 8 forbidden (e) 0 2 5 (f) 8 (g) 10 forbidden 2 (h) (i) 0 5 2 Figure 7.11. Example of the Fibonacci measurement algorithm The second step. For situation (a) we can use the next standard weight 3 and divide the line segment [5,13] by the IE in the Fibonacci ratio: 8=3+5. Two new situations (c) and (d) (Fig. 7.11) appear after the second step. Chapter 7 Algorithmic Measurement Theory 401 For the situation (b) the second step is “empty” because in accordance with restriction S it is forbidden to enclose the IE to the points of the line segment [0, 5] on the second step. The third step. For the situation (c) we can use the next standard weight 2 to divide the line segment [8,13] in the Fibonacci ratio: 5=3+2. Two new situations (f) and (g) appear after the third step. For the situation (e) we can return to the situation (b) on the third step. In accordance with the restriction S we can enclose the IE at any point of the line segment [0,5] on the third step. We can use the standard weight 2 to divide the line segment [0,5] in the Fibonacci ratio: 5=2+3. Two new situa tions (h) and (i) appear after the third step. It is easy to trace the actions of the Fibonacci algorithm for the next two steps (Fig. 7.11). We can see from this example that the essence of the Fibonacci measure ment algorithm is to divide the “indeterminacy interval” obtained on the pre ceding step in the Fibonacci ratio. It is easy to show that this general principle is valid for any arbitrary p. The division of the “indeterminacy interval” for this case is affected according to the recursive relation for Fibonacci pnumbers. 7.10. The Main Result of Algorithmic Measurement Theory 7.10.1. A Further Generalization of the BachetMendeleev Problem We can give a further generalization of the BachetMendeleev problem [20]. Suppose that we use k balances simultaneously (k is a natural number, 1, 2, 3 …). We can imagine that one and the same measurable weight Q is on the lefthand cups of all balances. Such a situation arises for the case of the “par allel measurement” of one and the same measured magnitude Q, when we compare the measurable magnitude Q with the standard weights by means of k “comparators” (such situation is widely used in the measurement of electric magnitudes). In this case the generalized BachetMendeleev problem can be formulated as follows: it is necessary to synthesize the optimal nstep mea surement algorithm to determine the value of the measurable magnitude Q with the help of k balances (“comparators”), which have the “inertness” р, for the condition that on each step the standard weights can be placed on the free cups of those balances, which are in the initial state. Alexey Stakhov THE MATHEMATICS OF HARMONY 402 Suppose that all balances, which participate in measuring, have the “inert ness” p (p=0, 1, 2, 3, …). This results in the problem to find the “parallel” optimal (n,k,S)algorithm, which uses k balances having the “inertness” p. For the simula tion of the “inertness” of the balances (or “Indicator Elements”) we can introduce the notion of the state of the jth IE on the lth step (j=1, 2, 3, …, k; l=1, 2, 3, …, n). Denote the latter by pj(l). It follows from the physical examination of the Bachet Mendeleev problem that the integer function pj(l) has the following properties: 0≤ pj(l)≤p (7.63) pj(l+1)= pj(l)1. (7.64) We can clarify the “physical sense” of the expressions (7.63) and (7.64). Note that the case pj(l)=0 means that the jth balance is in the initial position (Fig. 7.3 a) and the case pj(l)=p means that the balance is in the opposite position (Fig. 7.3 b). The expression (7.64) depicts the process of returning the jth balance from the opposite position (Fig. 7.3b) to the initial position (Fig. 7.3a). According to (7.64) the state of the IE is decreased by 1 on the next step. Thus, if the jth IE on the lth step of the algorithm turns out to be in the state pj(l)=p, then its states on the next steps will decrease successively according to the following: pj(l)=p→ pj(l+1)=p1→ pj(l+2)=p2→ … → pj(l+p1)=1→ pj(l+p)=0. Note that the case of pj(l)=0 means that the jth IE on the lth step of the algorithm may be enclosed to the points of the “indeterminacy interval.” After such preliminary remarks we can try to synthesize the optimal (n,k,S) algorithm. Before the lth step is carried out, we renumber all indicator ele ments so that their states would be arranged according to the rule: pk(l)≥ pk1(l)≥ pk2(l)≥ … ≥p2(l)≥ p1(l), (7.65) where pj(l) is the state of the jth IE on the lth step (j=1,2,3, …, k; l=1,2,3, …, n). We denote the IEstates on the first step of the (n,k,S)algorithm by p1, p2, p3, …, pk. The initial IEstates are disposed in accordance with (7.65), that is, pk≥ pk1≥ pk2≥…≥ p2≥ p1. (7.66) 7.10.2. The Main Recursive Relation of the Algorithmic Measurement Theory We denote the (n,k)exactness of the optimal (n,k,S)algorithm for this case by F(n,k)=Fp(n; p1, p2, p3, …, pk). Let the initial IEstates p1, p2, p3, …, pk be disposed according to (57) so that the first t states p1, p2, p3, …, pt are equal to 0, that is, Chapter 7 Algorithmic Measurement Theory 403 p1=p2= p3= …= pt =0. (7.67) This means that only the first t IE’s, which satisfy (7.66), can be enclosed to the points of the line segment AB on the first step of the (n, k, S)algorithm. Suppose there is the optimal (n, k, S)algorithm, which for the above restric , 0,..., 0 , pt +1 , pt + 2 , ..., pk equal parts of tion divides the line segment AB into Fp n; 0N t ∆=1. Let the first step of the algorithm be to enclose the t IE’s to the points C1, C2, …, Ct (Fig. 7.12). A C1 C2 Cj Cj+1 ... 1 2 Ct B ... j+1 t+1 Figure 7.12. Synthesis of the optimal (n, k, S)algorithm for the general case After the first step of the algorithm, the (t+1) situations X∈AC1, X∈ C1C2, …, X∈CjCj+1, …, X∈CtB appear dependent upon the “indications” of the t IE’s. Consider the situation X∈ΑC1. For this situation all t IE’s are on the right of the “unknown” point X and hence they turn into the states of p after the first step. In accordance with property (7.64) the rest of the (kt) IE’s decrease their states by 1 on the second step, that is, their states are: pt+11, pt+21, …, pk1. If we arrange all k IE’s in accordance with (7.65), we obtain the following IE states on the second step of the algorithm: pt +1 − 1, pt + 2 − 1, ..., pk − 1, p, p, ..., p . (7.68) t If we use the optimal (n1,k,S)algorithm with the initial IEstates (7.68) for this situation, then in accordance with the inductive assumption we can divide the line segment AC1 into F p n − 1; pt +1 − 1, pt + 2 − 1, ..., pk − 1, p, p, ..., p t equal parts of ∆ = 1. Consider the situation X∈C1C2. It follows from Fig. 7.12 that for this case the j IE’s are on the left of the point X and hence remain in the 0states, the (tj) IE’s turn into the p states and the rest of the (kt) IE’s decrease their states by 1, that is, pt+11, pt+21, …, pk1. Alexey Stakhov THE MATHEMATICS OF HARMONY 404 If we arrange all k IE’s in accordance with (7.65), we obtain the following sequence of the IEstates on the second step of the algorithm: 0, 0, ..., 0 , pt +1 − 1, pt + 2 − 1, ..., pk − 1, p, p, ..., p . (7.69) j t− j If we use the optimal (n1,k,S)algorithm with the initial IEstates (7.69) for this situation, then in accordance with the inductive assumption we divide the line segment CjCj+1 into F p n − 1; 0, 0, ..., 0 , pt +1 − 1, pt + 2 − 1, ..., pk − 1, p, p, ..., p j t− j equal parts of ∆ = 1. At least, it is easy to show that for situation X∈CtB the optimal (n1,k,S) algorithm divides the line segment CtB into F p n − 1; 0, 0, ..., 0 , pt +1 − 1, pt + 2 − 1, ..., pk − 1 t equal parts of ∆ = 1. Taking into consideration the evidence equality AB = AC1 + C1C2 + ... + C jC j +1 + ... + Ct B, we obtain the following recursive relation for the calculation of (n,k)exactness of the optimal (n,k,S)algorithm: F ( n, k ) = Fp n; 0, 0, ..., 0 , pt +1, pt +2 , ..., pk t (7.70) Fp n − 1; 0, 0, ..., 0 , pt +1 − 1, pt +2 − 1, ..., pk − 1, p, p, ..., p . j =0 j t− j Define the initial conditions for the calculation of Fp(n,k) according to the recursive relation (7.70). We take into consideration that for the cases n=1 and p1=p2=p3=…=pt=0, pt+1>0 the optimal (1,k,S)algorithm divides the line segment AB into t=1 equal parts, that is, F p 1; 0, 0, ..., 0 , pt +1 , pt + 2 , ..., pk = t + 1. (7.71) t Let us introduce the following definition: 0, with n < 0 F p ( n, k ) = . (7.72) 1, with n = 0 It follows from the “physical sense” of the function F(n,k)=Fp(n; p1, p2, p3, …, pk) that for the case n≥p1>0 we have: = t ∑ Chapter 7 Algorithmic Measurement Theory 405 Fp(n; p1, p2, p3, …, pk)=Fp(np1; 0, p2 p1, p3p1, …, pkp1). (7.73) The recursive relation (7.70) at the initial terms (7.71) and (7.72) is the main result of the algorithmic measurement theory. This recursive relation gives a the oretically infinite set of optimal measurement algorithms for the given n, k and p. 7.10.3. The Unusual Results The recursive relation (7.70) at the initial terms (7.71) and (7.72) contains a number of remarkable formulas of discrete mathematics. Consider now the case of p=0. This means that the IEstates are equal to 0 on each step of the algorithm, that is, (7.74) pk(l)= pk1(l)= … = p2(l)= p1(l)=0. Taking into consideration (7.74), we can prove that the recursive relation (7.70) and the initial condition (7.71) are reduced to the following: F p n; 0, 0, ..., 0 = ( k + 1) F p n − 1; 0, 0, ..., 0 (7.75) k k F p 1; 0, 0, ..., 0 = ( k + 1) . (7.76) k If we use the following definition F p n; 0, 0, ..., 0 = F ( n, k ) , k we can then represent the expressions (7.75) and (7.76) as follows: F(n,k)=(k+1)F(n1,k) (7.77) F(1,k)=k+1. (7.78) We can see that the expressions (7.77) and (7.78) coincide with the formu las (7.28) and (7.30) for the optimal (n,k,0)algorithm. As is shown above, the recursive function F(n,k) that is defined by the recursive relation (7.77) at the seeds (7.78) can be expressed in the explicit form given by the formula (7.31). It follows from this consideration that for the case p=0 the general opti mal (n,k,S)algorithm given by the formulas (7.70) and (7.71) is reduced to the optimal (n, k, 0)algorithm. Now, let us consider the case when p=∞ and p1=p2= p3= …= pk =0. (7.79) (7.80) Alexey Stakhov THE MATHEMATICS OF HARMONY 406 The “physical sense” of the conditions (7.78) and (7.80) are as follows. The condition (7.80) means that all k IE’s are in the 0states on the first step of the algorithm. According to the condition (7.79) every IE that is found on the right above the “unknown” point X at any step of the algorithm “leaves the field,” that is, cannot participate in further measurement. Let us consider the function F p n; 0, 0, ..., 0 , p, p, ..., p . (7.81) j k− j It follows from the “physical sense” that for the cases (7.79) and (7.80) the function (7.81) cannot depend on the (kj) IE’s that are found in the state p=∞. Hence, for this case the function (7.81) depends only on the number n of the algorithm steps and the number j of the IE’s that are found to the right of the point X on the first step of the algorithm. These “physical” considerations can be expressed in the following form: F p n; 0, 0, ..., 0 , p, p, ..., p = F ( n, j ) . (7.82) j k− j Note that F p n; p, p, ..., p = F ( n, 0 ) = 1. k (7.83) The “physical sense” of the expression (7.83) consists of the fact that all IE’s are found in the state p=∞ and we therefore do not have the possibility of de creasing the “indeterminacy interval.” Now let us return to formulas (7.70) and (7.71). It is clear that for the condi tions (7.79) and (7.80) the formulas (7.82) and (7.83) take the following forms, respectively: F p n; 0, 0, ..., 0 = F ( n, k ) (7.84) k F p 1; 0, 0, ..., 0 = k + 1. k (7.85) Taking into consideration definitions (7.82) and (7.83), we can represent the expressions (7.70) and (7.71) as follows: F(n,k)=1+ F(n1,k)+F(n1,1)+F(n1,2)+…+ F(n1,j)+…+ F(n1,k) (7.86) Chapter 7 Algorithmic Measurement Theory 407 F(1,k)=k+1. (7.87) Comparing the expressions (7.86) and (7.87) with the corresponding recur sive relations for the optimal (n,k,1)algorithm, we can see that these expres sions coincide and hence, F ( n, k ) = C nk+ k = C nn+ k . It follows from this consideration that for conditions (7.79) and (7.80) the general optimal (n,k,S)algorithm given by formulas (7.70) and (7.71) is reduced to the optimal (n,k,1)algorithm based on the arithmetical square. At least, consider the case k=1, p≥0 and p1=0. (7.88) This means that we deal with the (n,1,S)algorithm (k=1) and the IEstate on the first step of the algorithm equal to 0. Then, the expressions (7.70) and (7.71) for this case take the following form: Fp(n;0)= Fp(n1;0)+Fp(n1;p) (7.89) (7.90) Fp(1;0)=2. Let us consider the function Fp(n1;p) in (7.89). Using (7.73), we can write the following expression: Fp(n1;p)= Fp(np1;0). Then, the expression (7.89) can be written as follows: (7.91) Fp(n;0)= Fp(n1;0)+Fp(np1;0). Next let us introduce the following definition: (7.92) Fp(n;0)= Fp(n). Then, the expressions (7.92) and (7.90) can be written as follows: Fp(n)= Fp(n1)+Fp(np1) (7.93) Fp(1)=2. (7.94) Comparing the expressions (7.93) and (7.94) with the corresponding recur sive relations for the optimal Fibonacci measurement algorithms, we can see that these expressions coincide and, hence, for the case (7.88) the general optimal (n,k,S)algorithm given by the expressions (7.70) and (7.71) are reduced to the Fibonacci measurement algorithm. The main result of the algorithmic measurement theory together with all the unusual results is demonstrated in Fig. 7.13. Thus, the “unexpectedness” of the main result of the algorithmic measure ment theory (Fig. 7.13) consists of the following. The general recursive rela tion (7.70) at the seed (7.71) sets in general form an infinite number of new, Alexey Stakhov THE MATHEMATICS OF HARMONY 408 p=0 k ≥1 ( k + 1) 0≤ p≤∞ n ← ↓ k =1 2 n Binary numbers Fp ( n , k ) p=∞ → ↓ ← Fp ( n ) = Fp ( n − 1) + Fp ( n − 2 ) Fibonacci p − numbers C nk+ k = C nn+ k Binomial coefficients ↓ → n +1 Natural numbers Figure 7.13. The main result of the algorithmic measurement theory until now unknown, optimal measurement algorithms. All wellknown classical measurement algorithms that are used in measurement practice (the “binary” al gorithm, the “counting” algorithm, the “ruler” algorithm) are special limiting cas es of the optimal measurement algorithms. The main recursive relation (7.70) at the seed (7.71) includes a number of the wellknown combinatorial formulas as special cases, in particular, the formula (k+1)n, the formula C nk+ k = C nn+ k that gives binomial coefficients, the Fibonacci recursive relation, and finally the formulas for the “binary” (2n) and natural (n+1) numbers. 7.11. Mathematical Theories Isomorphic to Algorithmic Measurement Theory 7.11.1. What is an Isomorphism? The concept of Isomorphism [from the Greek words isos (equal), and morphe (shape)] is used widely in mathematics. Informally, isomorphism is a kind of correspondence between objects that show a relationship between two proper ties or operations. If there is isomorphism between two structures, we call these structures isomorphic. The word “isomorphism” applies when two complex structures can be reflected onto each other in such a way that each part of one structure fits a corresponding part of the other structure. We can give the following physical analogies of isomorphism: 1. A solid wood cube and a solid metallic cube are both solid cubes; although their physical nature differs, their geometric structures are isomorphic. 2. The Big Ben Clock in London and a wristwatch; although the clocks differ greatly in size, their timecounting mechanisms are isomorphic. Chapter 7 Algorithmic Measurement Theory 409 We can give different examples of isomorphism in mathematics. In abstract algebra, two basic isomorphisms are defined: Group Isomorphism, the isomorphism between groups, and Ring Isomorphism, the isomorphism between rings. There is also Graph Isomorphism in graph theory. In linear algebra, the isomorphism can also be defined as a Linear Map between two vector spaces. 7.11.2. Isomorphism between the “Balance” and “Rabbits” In the above we formulated the Asymmetry Principle of Measurement, the main methodological principle of algorithmic measurement theory. Note that this principle has practical origin because it reflects some essential properties of the comparators used in AnalogtoDigit Converters. For the first time, the isomorphism between bal ances and comparators was stated in this author’s Doctor of Science dissertation Syn thesis of Optimal Algorithms of AnalogtoDigit Conversion (1972). The above Asymmetry Principle of Measurement is based on the concept of the “inertial balances” used for measurement. If we define by I the initial posi tion of the balance (Fig. 7.3a) and by O the opposite position (Fig. 7.3b), then the functioning of the balance can be described by using two transitions: I I → (7.95) O (7.96) O → I. The transition (7.95) means that the balance can be in one of two extreme positions, I or O, after we place the next standard weight on the free cup of the balance. The transition (7.96) means that if the balance is in the position O, then during a certain time the balance is coming back into the initial position I. In Chapter 2, we described Fibonacci’s problem of rabbit reproduction. Let us recall that the “Law of rabbit reproduction” boils down to the follow ing rule. Each mature rabbit’s pair А gives birth to a newborn rabbit pair B during one month. The newborn rabbit’s pair becomes mature during one month and then in the following month said pair starts to give birth to one rabbit pair each month. Thus, the maturing of the newborn rabbits, that is, their transformation into a mature pair is performed in 1 month. We can model the process of “rabbit reproduction” by using two transitions: A → AB (7.97) (7.98) B → A. Note that the transition (7.97) simulates the process of the birth of the new born rabbit pair B and the transition (7.98) simulates the process of the maturing Alexey Stakhov THE MATHEMATICS OF HARMONY 410 of the newborn rabbit pair B. The transition (7.97) reflects an asymmetry of rabbit reproduction because the mature rabbit pair А is transformed into two nonidenti cal pairs, the mature rabbit pair А and the newborn rabbit pair B. Comparing the transitions (7.95) and (7.96) with the transitions (7.97) and (7.98), we can see analogies or isomorphism between them. Moreover, this isomorphism is confirmed by the fact that the solutions of these problems come to one and the same recursive numerical sequence, the Fibonacci numbers! 7.11.3. The Generalized “Asymmetry Principle” of Organic Nature Using the model of “rabbit reproduction,” which is described by the tran sitions (7.97) and (7.98), we can generalize the problem of rabbit reproduc tion in the following manner. Let us give a nonnegative integer p≥0 and formulate the following problem: “Let us suppose that in the enclosed place one pair of rabbits (female and male) is in the first day of January. This rabbit couple gives birth to a new pair of rabbits in the first day of February and then in the first day of each follow ing month. The newborn rabbit pair becomes mature in p months and then gives birth to a new rabbit pair each month thereafter. The question is: how many rabbit pairs will be in the enclosed place in one year, that is, in 12 months from the beginning of reproduction?” It is clear that for the case p=1 the generalized variant of the “rabbit repro duction” problem coincides with the classical “rabbit reproduction” problem described in Chapter 2. Note that the case p=0 corresponds to the idealized situation when the rabbits become mature at once after birth. One may model this case using the transition: A→AA. (7.99) It is clear that the transition (7.99) reflects symmetry of “rabbit reproduc tion” when the mature rabbit pair А turns into two identical mature rabbit pairs АA. It is easy to show that for this case the rabbits are reproduced according to the “Dichotomy principle,” that is, the amount of rabbits doubles each month: 1, 2, 4, 8, 16, 32, …. Now, let us consider the case p>0. We can analyze the above “rabbit reproduc tion” problem in greater detail, taking into consideration new conditions of “rabbit reproduction.” It is clear that the reproduction process is described by a more com plex system of transitions. Really, let A and B be the pairs of mature and newborn rabbits, respectively. Then the transition (7.97) simulates a process of the monthly appearance of the newborn pair В from each mature couple А. Chapter 7 Algorithmic Measurement Theory 411 Let us examine the process of transformation of the newborn pair В into the mature pair А. It is evident that during the maturation process the newborn pair В passes via the intermediate stages corresponding to each month: B → B1 B1 → B2 B2 → B3 (7.100) ... Bp −1 → A For example, for the case p=2 a process of transformation of the newborn pair into the mature pair is described by the following system of transitions: B → B1 (7.101) (7.102) B → A. Then, taking into consideration (7.97), (7.101) and (7.102), the process of “rabbit reproduction” for the case p=2 can be represented in Table 7.4. Table 7.4 The process of “rabbit reproduction” for the case p=2 Date Rabbit's pairs A B B1 A + B + B1 January, 1 A 1 0 0 1 February, 1 AB 1 1 0 2 March, 1 ABB1 1 1 1 3 April, 1 ABB1 A 2 1 1 4 May, 1 ABB1 AAB 3 2 1 6 June, 1 ABABB1 ABB1 AA 4 3 2 9 July, 1 ABB1 AABABB1 ABB1 A 6 4 3 13 Note that column А gives the number of the mature pairs for each stage of reproduction, column В gives the number of the newborn pairs, column B1 gives the number of the newborn pairs in stage B1, column A+B+B1 gives the general number of rabbit pairs for each stage of reproduction. The analysis of the numerical sequences in each column A : 1,1,1, 2, 3, 4, 6, 9,13,19,... B : 0,1,1,1, 2, 3, 4, 6, 9,13,19,... B1 : 0, 0,1,1,1, 2, 3, 4, 6, 9,13,19,... A + B + B1 : 1, 2, 3, 4, 6, 9,13,19,... demonstrates that they are subordinated to one and the same regularity: each number of the sequence is equal to the sum of the preceding number and the Alexey Stakhov THE MATHEMATICS OF HARMONY 412 number distant from the latter in 2 positions. However, as we know the Fibonacci 2numbers are subordinated to this regularity! Studying this process for the general case of р we come to the conclusion that the Fibonacci pnumbers are a solution to the generalized variant of the “rabbit reproduction” problem! They reflect the “Generalized Asymmetry Principle” of “Organic Nature.” At first appearance the above formulation of the generalized problem of “rabbit reproduction” appears to have no real physical sense. However, we should not hurry to such a conclusion! The article [171] is devoted to the appli cation of the generalized Fibonacci pnumbers for simulation of biological cell growth. The article affirms, “In kinetic analysis of cell growth, the assumption is usually made that cell division yields two daughter cells symmetrically. The essence of the semiconservative replication of chromosomal DNA implies com plete identity between daughter cells. Nonetheless, in bacteria, insects, nema todes, and plants, cell division is regularly asymmetric, with spatial and func tional differences between the two products of division…. Mechanism of asym metric division includes cytoplasmic and membrane localization of specific pro teins or of messenger RNA, differential methylation of the two strands of DNA in a chromosome, asymmetric segregation of centrioles and mitochondria, and bipolar differences in the spindle apparatus in mitosis.” In the models of cell growth based on the Fibonacci 2 and 3numbers are analyzed [171]. The authors of [171] made the following important conclusion: “Binary cell division is regularly asymmetric in most species. Growth by asymmetric binary division may be represented by the generalized Fibonacci equation …. Our models, for the first time at the single cell level, provide rational bases for the occurrence of Fibonacci and other recursive phyllotaxis and patterning in biology, founded on the occurrence of regular asymmetry of binary division.” Now we return to the Asymmetry Principle of Measurement. We can see that the transitions (7.100) for the generalized problem of “rabbit reproduc tion” are similar to the return of the balance to the initial position in the BashetMendeleev problem. This means that the generalized problem of “rab bit reproduction” is isomorphic to the Fibonacci measurement algorithm! We can generalize this idea. As the Asymmetry Principle of Measurement is the main idea of the algorithmic measurement theory, we can put forward the following hypothesis. The algorithmic measurement theory described by the very general recursive relation (7.70) and (7.71) is isomorphic to a General Theory of Biological Populations. This means that we can use the main result of the algorithmic measurement theory in the scientific field, which is very far from measurement theory. Chapter 7 Algorithmic Measurement Theory 413 7.11.4. Isomorphism between Measurement Algorithms and Positional Number Systems In the above we found that each measurement algorithm generates some posi tional number system. For example, the binary measuring algorithm generates the binary system, the optimal (n,k,0)algorithm generates all wellknown positional number systems, including the Babylonian sexagesimal system and the decimal sys tem. This means that the algorithmic measurement theory is isomorphic to the theory of positional number systems and we can interpret all optimal measurement algorithms as fundamentally new positional number systems. It is very important to emphasize, that the Algorithmic Measurement The ory [20, 21] as though precede a positional principle of number representa tion and positional numeral systems. This means that the creation of a gener al theory of positional numeral systems is reduced to a synthesis of the opti mal measurement algorithms. This remark is very important in order to de termine the role and place of the Algorithmic Measurement Theory in the de velopment of mathematics. In the Introduction, we attributed the Babylo nian positional principle of number representation to the greatest mathe matical discovery, which preceded number theory and mathematics. The Al gorithmic Measurement Theory infinitely enlarges a number of new, until now unknown positional numeral systems, turning mathematics back to the peri od of its origin. It is important to emphasize the “great practicability” of the Algorithmic Measurement Theory, which could become the basis of a new com puter arithmetic that is of fundamental interest to computer science. 7.12. Conclusion 1. It is well known that the “problem of measurement” played a fundamental role in mathematics. It stimulated the development of two basic theories of mathematics – geometry and number theory. The famous Bulgarian mathemati cian and academician L. Iliev wrote, “During the first epoch of mathematics development, from antiquity to the discovery of differential and integral calcu lus, mathematics, investigating first of all the measurement problems, created Euclidean geometry and number theory” [5]. That is why, since antiquity math ematical measurement theory together with number theory have been consid ered to be the fundamental mathematical theories underlying mathematics. Alexey Stakhov THE MATHEMATICS OF HARMONY 414 2. Let us compare the Classical Mathematical Measurement Theory [3, 4] with the Algorithmic Measurement Theory [20]. The creation of the classical measurement theory is connected with internal problems that appeared in math ematics after the discovery of Incommensurable Line Segments. This discovery staggered the Pythagoreans and caused the first crisis in the foundations of math ematics. The resulting Irrational Numbers became one of the fundamental concepts of mathematics. In order to overcome the first crisis in the founda tions of mathematics, the great mathematician Eudoxus (408 355 BC) sug gested the Method of Exhaustion, used by him for the creation of a Theory of Magnitudes, which preceded the Mathematical Measurement Theory complet ed in the 19th century. The interest in mathematical measurement theory in the 19th century increased again. The measurement theory received a new impulse thanks to Cantor’s theory of infinite sets. The mathematical mea surement theory was constructed on the basis of the Continuity Axioms EudoxusArchimedes’ Axiom and Cantor’s Axiom, which underlie classical mathematical measurement theory. Thanks to Cantor’s axiom, Cantor’s idea of Actual Infinity was introduced into mathematical measurement theory that allowed for the proof of the Basic Measurement Equality Q=qV. This equality for the given measurement unit V sets onetoone correspondence between geometric magnitudes Q and real numbers q. However, a construc tive approach to mathematics foundations, based on the Potential Infinity concept, demands the elimination of Cantor’s axiom from the Continuity Axioms, and puts forward a problem to prove the equality Q=qV from the constructive point of view (Constructive Measurement Theory). 3. In contrast to the classical mathematical measurement theory, the cre ation of the Algorithmic Measurement Theory is connected with the practical need for new algorithms of measurement to appear in theoretical metrology, in particular, in the theory of analogtodigital conversion [20]. It was found, that this practical problem is deeply connected with the ancient “Problem about the Choice of the Best Weights System.” This problem appeared in the 1202 work Liber Abaci by Fibonacci and is the first optimization problem in measurement theory. In the Russian mathematical literature, this problem is called the BashetMendeleev Problem [169]. The analysis of the BashetMen deleev problem from point of view of analogtodigital conversion resulted in the discovery of the very unusual measurement property known as the Asym metry Principle of Measurement, which in turn underlies Algorithmic Mea surement Theory [20, 21, 88, 90, 91]. The algorithmic measurement theory is an absolutely new and original mathematical theory. The synthesis of opti mal measurement algorithms is its main topic. This resulted in the discovery Chapter 7 Algorithmic Measurement Theory 415 of the infinite number of new, unknown until now, optimal measurement algo rithms, which have great theoretical and practical interest. These optimal measurement algorithms contain all the classical measurement algorithms (“binary,” “counting” and ruler” algorithms) and generate a wide class of new original measurement algorithms based on the Fibonacci pnumbers, binomi al coefficients, Pascal’s triangle, and so on. The main recursive relation of the algorithmic measurement theory (7.70), which generates an infinite number of different numerical sequences including natural numbers, binary numbers, Fibonacci pnumbers and binomial coefficients, is of fundamental interest for combinatorial analysis. 4. The idea of isomorphism resulted in an expansion of the applications of the algorithmic measurement theory. There is an isomorphism between measurement algorithms, the Fibonacci “rabbit reproduction,” and positional number systems. This idea gives us the right to put forward a new hypothesis that the new theory of measurement is the source of the three isomorphic theories, namely Algorithmic Measurement Theory, New Theory of Positional Number Systems and New Theory of Biological Populations. 5. The new theory of positional number systems returns mathematics to the Babylonian “positional principle of number representation,” which is consid ered to be the greatest mathematical achievement of Babylonian mathemat ics. We can emphasize that positional number systems were never considered in mathematics as a subject of serious mathematical research. That is why, a new theory of positional number systems, which follows from the algorith mic measurement theory, fills this gap, and a new theory of positional number systems should become a fundamental part of number theory. 6. We have a new theory of biological populations based upon the isomor phism between measurement algorithms and rabbit reproduction. In partic ular, as formulated above, the Generalized “Asymmetry Principle” of Organic Nature can play fundamental role here. This idea has been confirmed by con temporary research in the field of cell division [171]. Alexey Stakhov THE MATHEMATICS OF HARMONY 416 Chapter 8 Fibonacci Computers 8.1. A History of Computers 8.1.1. Era of Information An interesting view of the history of computer science is given in the book [172]. The time when people started creating instruments for hunting and man ual labor, is usually considered to be the beginning of human civilization. The secret of the prometheanlike “fire abduction” got lost in antiquity. However, the subsequent history of technical progress – from the “fire abduction” up to the discovery of nuclear energy – is the history of more and more powerful forces of nature being subjugated by humans: taming of animals, windmills, watermills, thermal engines, and atomic energy. During several millennia the main problem of material culture of humanity was to increase the muscular force of humans by various tools and machines. On the other hand, from aba cus stones up to modern computers, the efforts to create the tools strengthen ing human information processing mark the way for the accumulation of ideas in the general stream of scientific and technical progress, evident in the thin trickle of facts and museum pieces. Already in the earliest stages of the collective work development, a per son required some encoded signals for communication in order to synchro nize labor operations with others. Human speech, as the predecessor of infor mation technology and as the first transmitter of human knowledge, arose as a result of the complication of these information signals. At first, human knowl edge was accumulated in the form of oral stories and legends transmitted from one generation to the next. For the first time, the natural potential of a person accumulating and transmitting knowledge obtained “technological support” with the creation of Written Language. Written language became the first historical stage in the development of information technology. Chapter 8 Fibonacci Computers 417 We can find the proof of this stage imbedded deep in human history: such as, cave paintings (pictographs) and carvings (petroglyphs) in the form of people and animal images in stone, created twenty to twentyfive thousand years ago; a lunar calendar, engraved on bone more than twenty thousand years ago. However, ac cording to modern archaeological data, the time between the use of the first work tools (axe, trap, etc.) and tools for representing images (in stone, bone, etc.) is rough ly a million years. That is, nearly 99% of human history deals only with the develop ment of material labor tools. It appears that only the last (roughly) 1% of human history involves the development of information representation through images. The wheel, a recognized symbol of contemporary technical progress, was sup posedly invented in the Ancient East about 4 millennia BC, but the most an cient written code of laws, the first monument of the written language, arose only one thousand years later. About four millennia passed before the creation of the first Printing Machine in the middle of 15th century proclaimed to the world the onset of the Era of Printed Knowledge. The printing press rapidly increased the circulation of books as a passive means of information transmission. Printing machines for the first time created information accelerating the growth of pro ductive forces. The creation of Computers became the next stage in the develop ment of human civilization. These new machines for information processing ap peared in the middle of the 20th century, when the growing load of information problems became one of the most noticeable factors, limiting economic growth in industrially developed countries. Thus, during the history of human civilization before the 20th century, ma terial objects were the main subject of labor, and the economic power of countries was measured by material resources. At the end of the 20th century for the first time in the history of humanity, information became the main sub ject of labor in industrially developed countries. The tendency for a steady in crease in the information sphere in comparison to the material sphere is the most noticeable symptom of the “information era” approach. 8.1.2. The Basic Milestones in Computer Progress The history of computers, that is, the machines created for automation of information processing, is considered one of the least developed themes in the history of science. In the present book we will use a classification of periods of the computer development suggested in the book [173]. According to [173], the history of computers is divided into two main periods: 1. From abacus up to electronic computers 2. Electronic computers onward Alexey Stakhov THE MATHEMATICS OF HARMONY 418 8.1.2.1. Abacus The necessity for calculation arose in the early stages of human civilization: notches on bones and stones representing calculations were found beginning in the Paleolithic and Neolithic periods. The abacus was probably the first calculating device. The word “abacus” originates from the Greek word abax, which in turn comes from the Hebrew word avak (dust). The history of science testifies to the fact that many different cultures used the abacus – Babylon, China, India, Egypt, Arabian countries, Japan, Pakistan, and others. In medieval times, the word “abacus” was used as a synonym for mathematics. This was the meaning contained in the title of the mathematical book, Liber Abaci, written in 1202 by the famous Italian mathematician Leonardo from Pisa (Fibonacci). His book could be considered the original math ematical encyclopaedia of the Middle Ages. It is important to emphasize that in this early period of scientific development, mathematical history is inseparably linked to the history of primitive computers (abaci), because the problem of counting and calculation hastened the development of many important mathematical concepts. Note that at this stage, the Babylonian mathematicians discovered the positional principle of number representation underlying modern computers. The majority of the historians of mathematics emphasize that the creation of the first primitive abacus played a large role in the development of mathematics, and resulted in the development of a concept of natural number – one of our fundamental mathematical concepts. 8.1.2.2. Mechanical Calculation Machines Already philosophers in the Middle Ages advanced the problem of replacing human brain functions with various mechanisms. Several scientists were interest ed in the idea of the construction of “thinking” gear, and this idea stimulated the designs of the first mechanical calculating machines. In 1623, English professor of mathematics Schickard, a friend of Kepler, made such an attempt. In Kepler’s ar chive we find Schickard’s letter addressed to him. Schickard informed Kepler about a calculation machine he had constructed. Schickard’s machine was known, ap parently, by a very small group of people. Because of this limited exposure, for a long time it was thought that the first mechanical calculator was invented in 1642 by the famous French mathematician and physicist Pascal. About 50 copies of Pascal’s calculator were purportedly produced. Some copies of Pascal’s calculator are still available today. Five copies are in the Parisian Museum and one is in the Dresden physical and mathematical salon, which this author personally saw in Chapter 8 Fibonacci Computers 419 1988. In addition, the German mathematician Gottfried Leibniz, Wurttemberg’s pastor Gan, and the famous Russian mathematician and mechanician Chebyshev were inventors of original mechanical calculating machines. Production of calculating machines for the first time was started by Karl Thomas, the founder and chief of two Parisian insurance companies (“Fenix” and “Soleil”). In 1820, he built the calculation machine called Calculator. In 1821, 15 calculators were made in Thomas’ workshop. Later their production was in creased to 100 per year. These calculators had been produced during 100 years and had (certainly for that time) very impressive technical characteristics – two 8digit numbers could be multiplied in 15 seconds, and a 16digit number could be divided by the 8digit number in 25 seconds. A further improvement of calculators is connected with the name of Russian engineer Odner. Beginning in the 1890s, the triumphal procession of Odner’s calculators began with the serial production of 500 calculators per year in St. Petersburg, Russia. In the 20th century, Odner’s calculators were produced un der a variety of names in different countries. In the first quarter of the 20th cen tury, Odner’s calculators were the main mathematical machines used in many fields of human activity. In 1914, in Russia alone more than twentytwo thou sand Odnercalculators were produced. 8.1.2.3. Babbage’s Analytical Machine Charles Babbage was the mathematician, mechanical engineer and com puter scientist who originated the idea of a Programmable Computer. Babbage’s analytical machine was the best of the mechanical calculation machines. For the first time this machine embodies the idea of a completely automatic computer with program control. This idea was a revolutionary discovery in computer sci ence. It started to be used only in the middle of the 20th century in the design of the first electronic computers. The first copy of his analytical machine was made by Babbage in September 1834. From 1842 to 1848, Babbage paid significant attention to the design of his analytical machine. According to his idea, the machine should include three main blocks: 1. A device where digital information could be stored on sprocket registers. In modern computers a similar block is named the Stor age Device or Memory. Babbage named this part of the machine the Storehouse. Charles Babbage (1791 –1871) Alexey Stakhov THE MATHEMATICS OF HARMONY 420 2. A device where various operations with numbers, taken from the store house, can be carried out. Today this part of the machine is called the Arith metical Device. Babbage named this part of the machine the Factory. 3. A device for control of the operating sequence. Babbage didn’t have a spe cial name for this part of the machine. Today a similar computer block is called the Control Unit. In addition, his analytical machine had both Input and Output Devices. Babbage was the creator of the greatest invention in computer history, the Programmable Computer. Babbage’s main idea was only embodied in modern computers. A simple list of problems that was advanced by Babbage in his analytical machine is astonishing for its depth and foresight regarding the progress of modern computers. 8.1.2.4. Hollerith’s Tabulators The development of the theory of electricity and the theory of weak electri cal currents naturally resulted in the idea of using these currents in calculation devices. At first, the electric energy in calculation machines was used only as a motion force for the actuation of the calculator mechanism in place of the hand. Such machines were called Electromechanical Machines. At the end of the 19th century, the need for population census processing arose. In 1888, an employee of the U.S. Census Bureau, Herman Hollerith, designed a calculation machine that was used for processing the population census. Hollerith named his machine the Tabulator. His tabulator was intended to automatically process punch cards for the population census. In his machine Hollerith applied the benefits of weak current engineering and his machine was based on the electro mechanical principle. The same electromechanical principle was used for improve ment of the technical parameters of calculators. Thus, at the beginning of the 20th century, digital computer engineering developed in two directions. The first di rection was the small Calculators intended for the mechanization of the elementa ry arithmetical operations. The second direction was that of Tabulators intended for the processing of statistical information. 8.1.2.5. Electromechanical Computers by Zuse, Aiken and Stibitz In the 1930s and 1940s the development of tabulators, as well as, the achieve ments in the field of the electromechanical relay applications (for example, in automatic telephone stations) resulted in the development of computer machines similar to Babbage’s machine by their structure. The first universal electrome chanical computers with programmed control were designed in Germany and the U.S. in the early 1940s. German engineer Konrad Zuse, along with American Chapter 8 Fibonacci Computers 421 scientists Howard Aiken from Harvard University and George Robert Stibitz from Bell Labs, played significant roles in these projects. Zuse developed several models of electromechanical computers. The first model Z1 was constructed over a period of two years (19361938). It was based completely on mechanical elements, rendering it somewhat unsatisfactory. The next models Z2, Z3, Z4 used electromagnetic relays as their basis. An elabora tion of the Z3 machine was completed in 1941. This machine was the first uni versal computer with programmed control. The computer performed eight com mands including four arithmetical operations, multiplying by negative numbers, and calculating square roots. All calculations were fulfilled in the binary system using a floating point. Aiken’s universal digital computer used standard punch cards produced by IBM (USA). In August, 1944 the project was completed, and the machine called MARC1 was installed at Harvard University, where the machine was used thereafter for over 15 years. If we compare the machine MARC1 with the machine Z3, we should note that Zuse’s machine exceeded Aiken’s machine from the point of view of its circuit and structural solutions. Also the works of American scientist Stibitz played an essential role in the creation of the first computers, which were intended for the fulfilment of complicated scientific and technical calculations. In 1938, working for Bell Laboratory, Stibitz developed the computer “Bell1” capable of operating with complex numbers. In 1947, Stibitz developed a highpower universal computer with programmed control (the machine “Bell5”) based on elec tromagnetic relays. At the same time, the computing laboratory of Harvard University developed the large programcontrolled computer MARC2 which is also based on electromagnetic relays. 8.1.2.6. ENIAC One of the first attempts to use electronic elements in computers was launched in the U.S. in 19391941 at the State College of Iowa (now the Iowa State University). Physics professor John Atanasov was the principal developer of this project. Unfortunately, the project was not completed, but in 1973 the U.S. Federal Court confirmed Atanasov’s priority in the design of the first elec tronic computer project. Work on the creation of the electronic computer ENIAC was started in 1942, at Pennsylvania University under John Mauchly and John Eckert’s leadership. This works was completed at the end of 1945, and in February 1946 the ma chine’s first public demonstration was held. The important role of ENIAC in the development of computer science is clear from the fact that it was the first ma Alexey Stakhov THE MATHEMATICS OF HARMONY 422 chine in which electronic elements were used for the realization of arithmetical and logical operations, and also the storage of information. The use of new elec tronic technology in this machine allowed for an increase of computer operation speed by approximately 1000 times greater in comparison with electromechan ical computers. 8.1.2.7. John von Neumann’s Principles The ENIAC confirmed in practice the high efficiency of electronic technolo gy in computers. The problem of maximal real ization of the large advantages of electronic tech nology arose for computer designers. It was nec essary to analyze the strengths and weaknesses of the ENIAC project and to give appropriate rec ommendations. A brilliant solution to this prob lem emerged in the famous report Preliminary Dis cussion on Logic Designing the Electronic Com puting Device (1946). This report, written by famed mathematician John von Neumann and his colleagues from Princeton University, Goldstein and Berks, presented the project of developing the John von Neumann (19031957) new electronic computer. The essence of the primary recommendations of the report consisted of the following: 1. The machines based on electronic elements should work in Binary System instead of the decimal system. 2. The program should be stored in the machine block called the Storage Device, which should have sufficient capacity and appropriate speeds for ac cess and entry of a variety of program commands. 3. Programs, as well as numbers, with which the machine operates, should be represented in Binary Code. Thus, Commands and Numbers Should Have One and the Same Form of Representation. This meant that all programs and all intermediate outcomes of calculations, constants and other numbers should be stored in one and the same storage device. 4. Difficulties of physical realization of the storage device, speed of which should correspond to the speed of logical elements, demanded a Hierarchical Organization of Memory. 5. The arithmetical device of the machine should be constructed on the basis of the Logical Summation Element; it is not recommended to design special devices for the realization of other arithmetical operations. Chapter 8 Fibonacci Computers 423 6. The machine should use the Parallel Principle of Organization of Com puting Processes, that is, the operations on binary words should be performed on all digits simultaneously. John von Neumann’s Principles, in general, were some of the most important contributions to the development of universal electronic computers. The further development of electronic computers was based upon these principles. The changes concerned only the technology of the element base. Depending on the electronic base used, electronic computers started developing in the following three directions: 1. Computers based on the electronic lamps. 2. Computers based on the discrete semiconducting and magnetic elements. 3. Computers based on the integrated schemes. 8.1.2.8. The Phenomenon of Personal Calculators The capacity of the present book, which is not a specialty textbook on the history of the computer, does not allow us to refer to many achievements of mod ern computers. We can only discuss one phenomenon of modern computer his tory – that of personal computers (PC), which were the first mass tools for the active formalization of professional knowledge. Varied influence of the PC on progress of industrial society is comparable to book printing, which began the information era. Continuing this analogy: while a book was and remains the tool for mass duplication and passive storage of knowledge, the PC is the first mass produced tool for direct transformation of professional knowledge into an active industrial product. The discovery of the PC phenomenon in the U.S. was con nected with the name Steve Jobs, founder of Apple Computer. In 1980, Steve Jobs defined this type of computer as the tool for increasing the natural potentialities of human intellect. The PC, more than any other tool for infor mation processing, brings forward the information era. According to the opinion of many competent scientists, we can now identify a few of the symptoms, which testi fy to the beginning of the transition within the industrially developed countries to a qualitatively new stage of technological development – the information era: 1. Information becomes the main subject of labor in public production of the industrially developed countries. 2. The time of doubling accumulated scientific knowledge is something like 23 years. 3. The material expenses for storage, transmission and processing of infor mation begin to exceed similar costs for energy expenses. 4. Humanity for the first time in its history becomes the really observed astro nomical “cosmic factor,” as the level of Earth radio emissions in the separate parts of the radio diapason approaches the level of the Sun’s radio emissions. Alexey Stakhov THE MATHEMATICS OF HARMONY 424 8.2. Basic Stages in the History of Numeral Systems The creation of the first numeral systems is attributed to the period of the mathematics origin, when the necessity to count things, and measure time, land and product quantities, resulted in the development of the basic principles of the arithmetic of natural numbers. In the history of numeral systems we can identify several stages: the Initial Stage of Counting, NonPositional Numeral Systems, Alphabetic Notations, and Positional Numeral Systems. Initially people used body parts, fingers, sticks, knots, etc. for representation of counted sets. Article [2] emphasizes that “despite the extreme primitiveness of this way of number representation, it played an exclusive role in the development of the number concept.” This statement confirms that numeral systems played an im portant role in the formation of the concept of natural numbers – one of mathe matics’ most fundamental concepts. 8.2.1. Babylonian Sexagecimal Numeral System At an early stage in the history of mathematics, one of the greatest mathemat ical discoveries was made. This was the Positional Principle of number representa tion. “The Babylonian sexagecimal numeral system that arose in approximately 2000 B.C. was the first numeral system based on the positional principle” [2]. There were a number of competing hypotheses explaining its origin. M. Cantor original ly assumed that Sumerians (the primary population of the Euphrates valley) con sidered that one year consisted of 360 days and therefore that base 60 had an as tronomical origin. According to Kevitch’s hypothesis, in the Euphrates valley two persons met. The first person knew the decimal system, and the other one knew the base 6 system (Kevitch explains the occurrence of such base by finger count ing, in which the compressed hand meant the number 6). After the two systems were merged the compromise base of 60=6×10 arose. We should note that Cantor’s and Kevitch’s hypotheses concerned the ques tion of the origin of base 60, but they do not explain the origin of the positional principle of number representation. Neugebauer’s hypothesis presented in the book [174] gives an answer to this last question. According to his hypothesis, the positional principle has a measurement origin. Neugebauer asserts [174] that “the basic stages of the origin of the positional principle in Babylon were the following: (1) determination of a quantitative ratio between two independent existing systems of measures and (2) omitting the names of numerals in the writing.” According to his opinion, these stages of the origin of the positional numeral systems had a common foundation. He emphasizes that Chapter 8 Fibonacci Computers 425 “the positional sexagecimal number system was the natural outcome of a long his torical development, which does not differ from similar processes in other cultures.” However, we can explain the choice of the number 60 as the base of the Baby lonian system from the point of view of the Universal Harmony concept. As is well known, the dodecahedron was the main geometric figure expressing the Universal Harmony in ancient science. As is shown in Chapter 3, the numerical parameters of the dodecahedron (12, 30, 60) were connected to the basic cycles of the Solar system (Jupiter’s 12year cycle, Saturn’s 30year cycle and the Solar system’s 60 year main cycle). The ancient scientists chose the numbers 12, 30, 60 and derived from them the number 360=12×30 as the “main numbers” of the calendar system and systems of time and angular values measurement. Taking into consideration the deep connection between the Babylonian and Egyptian cultures, we may ad vance the hypothesis that the base 60 of the Babylonian number system was cho sen by the Babylonians from astronomical considerations. This hypothesis coin cides with Cantor’s hypothesis about the astronomical origin of base 60. 8.2.2. Alphabetic and Roman Systems of Numeration Many people used the socalled alphabetic system of numeration, where nu merical values were attributed to the letters of the alphabet. For instance, ancient Greek, Hebrew, Sanskrit, Slavic and Arabic alphabets employed this idea. The Roman system of numeration is widely known. It uses special numerals for the representation of the “nodal” numbers (I, V, X, L, C, D, M). The greatest shortcoming of Roman numeration was the fact that it was not adapted to per form arithmetical operations in written form. 8.2.3. Mayan System The origin of positional numeral systems is considered to be one of the main events in the history of material culture. Many people took part in its creation. In the 6th century AD a similar numeral system arose within the Maya people. It is widely accepted that the number 20 is the base of the Mayan number sys tem and, therefore, may have a “finger/toe” origin (10+10=20). It is well known that the number system with base 20 used the following set of the “nodal” num bers for number representation: {1, 201, 202, 203}. However, the Mayan number system used another set of “nodal” numbers: {1, 20, 360, 20×360, 202×360,…}. This means that the next “nodal” number, which follows after 20, is equal to 360 (in stead of 202=400) and all subsequent “nodal” numbers are derivable from the numbers 20 and 360. As is emphasized in the article [2], this somewhat unusual numbering system can be explained by the fact that “the Maya divided one year Alexey Stakhov THE MATHEMATICS OF HARMONY 426 into 18 months, with 20 days in every month, plus five more days.” Thus, like the base of the Babylonian system, the “nodal” numbers of Mayan notation have an astronomical origin. It is essential to emphasize, that the structure of the Maya calendar year (360+5=18×20+5) is similar to the structure of the Egyptian cal endar year (360+5=12×30+5). 8.2.4. Decimal System We use a decimal system for daily calculations. In mathematics this numeral system is called the HinduArabic or Arabic numeral system. This number system was invented by Hindu mathematicians in approximately the 6th8th centuries AD and it is the predecessor of the traditional decimal system. This number sys tem uses the number 10 as the base and requires 10 different numerals, the digits 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 for the representation of numbers. From India, the decimal system penetrated into the Arabic world. Arabic mathematicians extended the system to decimal fractions. In 9th century Muhammad ibn Musa alKhwarizmi wrote an important work about the decimal system. Thanks to a translation of al Khwarizmi’s work, the decimal system began to be used in Europe. There are different opinions concerning the choice of the number 10 as the base of the decimal system; the most widespread opinion is that the base of 10 has a “finger” origin. However, it is necessary to remind oneself that the number 10 always had special significance in ancient science. Pythagoreans considered this number to be the “Sacred Number” and named it the Tetractys. The number 10=1+2+3+4 was considered by Pythagoreans to be one of the greatest values, being “a symbol of the Universe,” because it comprised four “basic elements”: the One or Monad symbolizes spirit, out of which all the visible world appears; the Two or Dyad (2=1+1), symbolizes the material atom; the Three or Triad (3=2+1), symbolizes the living world. The Four or Tetrad (4=3+1) connected the living world with the Monad and consequently symbolizes the Whole, that is, the Visible and Invisible World together. As the Tetractys 10=1+2+3+4, which meant that the Tetractys, 10, expressed Everything. Thus, the hypothesis about the “harmonic” origin of base 10 is as tenable as the hypothesis of the “finger” origin of the decimal system. 8.2.5. Binary System The binary system uses two symbols, 1 and 0, for number representation. The number 2 is its base or radix. In connection with the development of com puter technology, the binary system was introduced into modern science. Dis cussing the history of the binary system, we should note that the ancient Indian Chapter 8 Fibonacci Computers 427 writer and mathematician Pingala pre sented in 2nd century AD the first known description of a binary system, more than 1500 years before its discov ery by German mathematician Got tfried Leibniz in 1695. Pingala’s work also contains elements of combinatorics (Pascal triangle) and the basic ideas of Fibonacci numbers. The modern binary system was ful ly documented by Gottfried Leibniz in 17th century in his article Explication de l’Arithmйtique Binaire. Leibniz’s sys tem used 0 and 1 in a manner similar to the modern binary system. Gottfried Wilhelm von Leibniz (1646 1716) was Gottfried Wilhelm von Leibniz a German polymath who wrote mostly (16461716) in French and Latin. He invented calcu lus independently of Newton. Leibniz also made major contributions to physics and technology, and anticipated notions that surfaced much later in biology, med icine, geology, probability theory, psychology, and information science. His con tributions to this vast array of subjects are scattered in journals and in numerous letters and unpublished manuscripts. To date, there is no complete edition of Leib niz’s writings, and a full list of his accomplishments is not yet known. The binary arithmetic developed by Leibnitz is much similar to arithmetic in other numeral systems. Summation, subtraction, multiplication, and division can be performed on binary numerals. All arithmetical operations in the binary system are reduced to binary summation. The summation of two singledigit binary numbers is relatively simple: 0 + 0 = 0 0 + 1 = 1 1 + 0 = 1 1 + 1 = 10 (carry of 1) In 1854, British mathematician George Boole published an important paper detailing a system of logic that became known as Boolean Algebra. His logic system became the instrumental tool for the development of the binary system, particularly in its implementation in electronic circuits. In 1937, Claude Shannon presented his master’s thesis at MIT that initiated the implementation of Boolean algebra and binary arithmetic using electronic Alexey Stakhov THE MATHEMATICS OF HARMONY 428 relays and switches for the first time in history. Entitled A Symbolic Analysis of Relay and Switching Circuits, Shannon’s thesis essentially developed the practi cal digital circuit design. In 1946, John von Neumann, Goldstein and Berks in their Preliminary Discussion on Logic Designing the Electronic Computing Device gave a decisive preference for the use of the binary system in electronic computers. 8.2.6. Exotic Number Systems However, we should note that the binary system suffers from some serious disadvantages [175, 176]: 1. The problem of a number sign. In the binary system we can represent in the “direct” code only positive numbers. That is why, for the representation of negative numbers we need to use special codes additional and inverse codes. This complicates the arithmetical structures of computers and influ ences the internal performance of a computer. 2. The problem of error detection. As is well known, all computer devices are subjected to severe external and internal influences that are a cause of vari ous errors, which can appear in the information transmission and process ing. Unfortunately, such errors cannot be detected because all binary code combinations are permissible. 3. The problem of long carryover. If, for example, we sum the two binary numbers 011111111+000000001=100000000, we should take into consider ation the “long carryover” from the lowest digit to the highest digit appear ing in such a summation. This “long carryover” fundamentally influences the internal performance of computers. The intention to overcome these disadvantages of the binary system led to the appearance of different numeral systems with “exotic” titles and properties: Sys tem for Residual Classes, Ternary Symmetrical Numeral System, Numeral System with Complex Radix, Negapositional, Factorial, Binomial Numeral Systems [175, 176], etc. All of them had certain advantages in comparison with the binary sys tem, and were directed at the improvement of computer characteristics; some of them became the basis for the creation of new computer projects (e.g. the ternary computer “Setun,” the computer based on system for residual classes, and so on). However, there is also another interesting aspect to this problem. Four mil lennia after the invention by Babylonians of the positional principle of number representation, we observe some kind of “Renaissance” in the field of numeral systems. Thanks to the efforts of the experts in computers and mathematics, it is as if we once again have returned to the origin of computation, when the numeral Chapter 8 Fibonacci Computers 429 systems defined the subjectmatter and were the heart of mathematics (Baby lon, Ancient Egypt, India, and China). The main purpose of chapters 8, 9 and 10 is to demonstrate that new meth ods of number representation based on the golden mean and Fibonacci numbers can become the source of a new and unusual computer arithmetic and from this will follow new and exciting computer projects. 8.3. Fibonacci pCodes 8.3.1. Zeckendorf Sums Many numbertheorists know about Zeckendorf sums [16], however, few know about the man for whom these sums are named. The Fibonacci Association con siders Edouard Zeckendorf to be one of the famous Fibonacci mathematicians of the 20th century although he did not have mathematical education. Zeckendorf was a medical doctor and was licensed for dental surgery. Mathematics was Zeck endorf’s hobby. Through the years, he published several mathematical articles. The most important article, on Zeckendorf sums, was published in 1939. In it, Zeckendorf proved that each positive integer can be represented as the unique sum of nonadjacent Fibonacci numbers, as exemplified below: 38=34+3+1; 39=34+5; 40=34+5+1; 41=34+5+2; 42=34+8. The numerous articles published in The Fibonacci Quarterly discussed Zeck endorf sums and their generalizations. Let us write Zeckendorf representation in the following form: N=anFn+an1Fn1+…+aiFi+…+a1F1, (8.1) where ai ∈{0,1} is the binary numeral of the ith digit of the representation (8.1); n is the number of digits of the representation (8.1); Fi is a Fibonacci number given by the recursive relation Fn= Fn1+Fn+2 at the seeds (8.2) (8.3) F1= F2=1. It is important to emphasize that the Fibonacci numbers Fi (i=1,2,3,…,n) are used as the digit weights in Zekendorf’s representation (8.1). The binary representation of Zeckendorf sum (8.1) has the following form: (8.4) N = an an1 ... ai ... a1, where ai ∈{0,1} is a binary numeral of the ith digit of the binary representation Alexey Stakhov THE MATHEMATICS OF HARMONY 430 (8.4). Thus, Zeckendorf sum (8.1) can be considered as the binary representa tion (8.4) with “Fibonacci weights.” It is clear that the minimal number Nmin that can be represented by using Zeckendorf sum (8.1) is equal to 0 and has the following binary representation: N min = 0 = 00 ...0 . N (8.5) n The maximal number Nmax that can be represented by using Zeckendorf sum (8.1) has the following binary representation: N max = 11 ...1. N (8.6) n The binary representation (8.6) is an abridged representation of the following sum: Nmax=Fn+Fn1+…+Fi+…+F1 Using the formula (2.20), we can write: (8.7) (8.8) Nmax=Fn+21. This means that we have proved the following theorem. Theorem 8.1. Using the ndigit Zeckendorf sum (8.1), we can represent Fn+2 integers in the range from 0 to Fn+21. 8.3.2. Definition of the Fibonacci pCode In Chapter 7 we developed the socalled Algorithmic Measurement Theory [20, 21] that can be a source of new ideas in numeral systems field. Fibonacci codes and Fibonacci arithmetic, which are a wide generalization of the binary system, are amongst the most interesting applications of the algorithmic mea surement theory. The Fibonacci measurement algorithms, which were examined in Chapter 7, are isomorphic to the following sum: (8.9) N=anFp(n)+an1Fp(n1)+…+aiFp(i)+…+a1Fp(1), where N is a natural number, ai ∈{0,1} is a binary numeral of the ith digit of the code (8.9); n is the digit number of the code (8.9); the Fibonacci number Fp(i) is the ith digit weight calculated in accordance with the recursive relation (4.18) at the seeds (4.19). The positional representation of the natural number N in the form (8.9) is called a Fibonacci pcode [20]. The abridged representation of the Fibonacci pcode (8.9) has the same form (8.4) as the abridged representation of Zecken dorf sum (8.1). Note that the notion of a Fibonacci pcode (8.9) includes an infinite number of different positional “binary” representations of natural numbers because ev Chapter 8 Fibonacci Computers 431 ery p produces its own Fibonacci pcode (p=0,1,2,3,…). In particular, for the case p=0 the Fibonacci pcode (8.9) is reduced to the classical binary code: (8.10) N=an2n1+an12n2+…+ai2i1+…+a120 . For the case p=1 the Fibonacci pcode (8.9) is reduced to Zeckendorf sum (8.1). Now, let us consider the partial case p=∞. For this case each Fibonacci p number is equal to 1, that is, for any integer i we have: Fp(i)=1. Then the sum (8.9) takes the form of the “unitary code”: N = 1+ 1 + ... + 1. (8.11) Thus, the Fibonacci pcode given by (8.9) is a very wide generalization of the binary system (8.10) and Zeckendorf sum (8.1) that are partial cases of the Fibonacci pcode (8.9) for the cases p=0 and p=1, respectively. On the other hand, the Fibonacci pcode (8.9) includes the socalled “unitary code” (8.11) as another extreme case for p=∞. N 8.3.3. The Range of Number Representation in the Fibonacci pCode Consider the set of ndigit binary words. The number of them is equal to 2n. For the classical binary code (8.10) (p=0) the mapping of the set of ndigit bina ry words onto the set of natural numbers has the following peculiarities: (a) Uniqueness of mapping. This means that for the infinite n there are one toone correspondences between natural numbers and sums (8.10), that is, every integer N has only one representation in the form (8.10). (b) For a given n with the help of the binary code (8.10) we can represent all integers in the range from 0 to 2n 1, that is, the range of number representa tion is equal to 2n. (c) The minimal number 0 and the maximal number 2n1 have the following binary representations in the binary code (8.10), respectively: N min = 0 = 00 ...0 N n N max = 2n − 1 = 11 ...1. N n For the Fibonacci pcode (8.9) the mapping of the ndigit binary words onto natural numbers has distinct peculiarities for the case p>0. Let n=5. Then, for the cases p=1 and p=2 the mappings of the 5digit Fi bonacci pcode (8.9) onto the natural numbers have the forms, represented in Tables 8.1 and 8.2, respectively. Alexey Stakhov THE MATHEMATICS OF HARMONY 432 Table 8.1. Mapping the Fibonacci 1code onto natural numbers An analysis of Tables 8.1 and 8.2 leads to the following CC 5 3 2 1 1 N CC 5 3 2 1 1 N peculiarities of binary represen A0 0 0 0 0 0 0 A16 1 0 0 0 0 5 tations of natural numbers in A1 0 0 0 0 1 1 A17 1 0 0 0 1 6 Fibonacci pcodes (8.9). By us A2 0 0 0 1 0 1 A18 1 0 0 1 0 6 ing the 5digit Fibonacci 1code A3 0 0 0 1 1 2 A19 1 0 0 1 1 7 (Table 8.1), we can represent 13 A4 0 0 1 0 0 2 A20 1 0 1 0 0 7 integers in the range from 0 to A5 0 0 1 0 1 3 A21 1 0 1 0 1 8 12, inclusively. Note that the A6 0 0 1 1 0 3 A22 1 0 1 1 0 8 number 13 is the Fibonacci 1 A7 0 0 1 1 1 4 A23 1 0 1 1 1 9 number with index 7, that is, A8 0 1 0 0 0 3 A24 1 1 0 0 0 8 F1(7)= F7=13. We can see from A9 0 1 0 0 1 4 A25 1 1 0 0 1 9 Table 8.2 that by using the 5 A10 0 1 0 1 0 4 A26 1 1 0 1 0 9 digit Fibonacci 2code (Table A11 0 1 0 1 1 5 A27 1 1 0 1 1 10 8.2) we can represent 9 integers A12 0 1 1 0 0 5 A28 1 1 1 0 0 10 in the range from 0 to 8, inclu A13 0 1 1 0 1 6 A29 1 1 1 0 1 11 sively. Here, the number 9 is the A14 0 1 1 1 0 6 A30 1 1 1 1 0 11 Fibonacci 2number with index A15 0 1 1 1 1 7 A31 1 1 1 1 1 12 8, that is, F2(8)=9. The results of this consideration are partial cases of the following theorem. Theorem 8.2. For the giv Table 8.2. Mapping the Fibonacci 2code en integers n>0 and p>0, with onto natural numbers the help of the ndigit Fibonac CC 3 2 1 1 1 N CC 3 2 1 1 1 N ci pcode, we can represent A0 0 0 0 0 0 0 A16 1 0 0 0 0 3 F p (n+p+1) integers in the A1 0 0 0 0 1 1 A17 1 0 0 0 1 4 range from 0 to Fp(n+p+1)1, A2 0 0 0 1 0 1 A18 1 0 0 1 0 4 inclusively. A3 0 0 0 1 1 2 A19 1 0 0 1 1 5 Proof. It is clear that the A4 0 0 1 0 0 1 A20 1 0 1 0 0 4 minimal number N min can be A5 0 0 1 0 1 2 A21 1 0 1 0 1 5 represented in the Fibonacci p A6 0 0 1 1 0 2 A22 1 0 1 1 0 5 code (8.9) as follows: A7 0 0 1 1 1 3 A23 1 0 1 1 1 6 N min = 0 = 00 ...0 . N n (8.12) The binary representa tion of the maximal number Nmax has the following form: N max = 11 ...1. N n (8.13) A8 A9 A10 A11 A12 A13 A14 A15 0 1 0 0 1 0 0 1 0 0 1 0 0 1 1 0 1 1 0 1 1 0 1 1 0 0 2 0 1 1 0 0 1 1 3 3 4 3 4 4 5 1 0 1 0 1 0 1 A24 A25 A26 A27 A28 A29 A30 A31 1 1 0 0 0 5 1 1 1 1 1 1 1 6 6 7 6 7 7 8 1 1 1 1 1 1 1 0 0 0 1 1 1 1 0 1 1 0 0 1 1 1 0 1 0 1 0 1 Chapter 8 Fibonacci Computers 433 The binary combination (8.13) is an abridged representation of the fol lowing sum: Nmax=Fp(n)+ Fp(n1)+…+Fp(i)+…+Fp(1). Using the formula (4.39), we can write: (8.14) Nmax=Fp(n+p+1)1. (8.15) The theorem is proved. Note that for the case p=0, F0(n+1)=2n and Theorem 8.2 is reduced to the wellknown theorem about the range of number representation for the classical binary system. This range is equal to 2n for the ndigit binary code (8.10). 8.3.4. Multiplicity of Number Representation The Multiplicity of number representation in the form (8.9) is the next peculiarity of the Fibonacci p Table 8.3. Mapping natural numbers to binary code (8.9) for the case p>0. By combinations in the Fibonacci 1 and 2 codes accepting the minimal number p =1 p=2 0 given by (8.12) and the max 0 = { A0 } imal number given by (8.13), 1 = { A1 , A2 } the remaining integers from 2 = { A3 , A4 } 0 = { A0 } the range [0, F p(n+p+1)1] 1 = { A1 , A2 , A4 } 3 = { A5 , A6 , A8 } have more than one represen tation in the form (8.9). This 2 = { A3 , A5 , A6 , A8 } 4 = { A7 , A9 , A10 } means that all integers in the 3 = {A7 , A9 , A10 , A12 , A16 } 5 = { A11 , A12 , A16 } range [1, Fp(n+p+1)2] have 6 = { A13 , A14 , A17 , A18 } 4 = {A11 , A13 , A14 , A17 , A18 , A20 } multiple representations in the 7 = { A15 , A19 , A20 } 5 = { A15 , A19 , A21 , A22 , A24 } Fibonacci pcode (8.9) for the 8 = { A21 , A22 , A24 } 6 = {A23 , A25 , A26 , A28 } case p>0. 9 = { A23 , A25 , A26 } 7 = { A27 , A29 , A30 } Now, let us examine the 10 = { A27 , A28 } 8 = { A31 } mapping of integers onto the 5 digit binary code combinations 11 = { A29 , A30 } A in accordance with Table 8.1 12 = { A31 } (p=1) and Table 8.2 (p=2). This mapping is represented in Table 8.3. Note that for the arbitrary p the minimal and the maximal numbers have only binary representations in the ndigit Fibonacci pcode: 0 = 0 0 ... 0 (n digits) and Fp(n+p) 1 = 1 1 ... 1 (n digits). Alexey Stakhov THE MATHEMATICS OF HARMONY 434 8.4. Minimal Form and Redundancy of the Fibonacci pCode 8.4.1. Convolution and Devolution The different binary representations of one and the same integer N in the Fibonacci pcode (8.9) for the case p>0 may be obtained from one another by means of the peculiar code transformations called Convolution and Devolution of the binary digits. These code transformations are carried out within the scope of one binary combination and follow from the basic recursive relation (4.18) that connects adjacent digit weights of the Fibonacci pcode (8.9). The idea of such code transformations consists in the following. Let us consider the abridged representation (8.4) of integer N in the Fibonacci pcode (8.9). Suppose that the binary numerals of the lth, (l1)th and (lp1) th digits in (8.4) are equal to 0, 1, 1, respectively, that is, N=an an1 … al+1 01 al2… alp 1 alp2… a1. (8.16) This binary representation (8.16) may be transformed into another binary representation of the same number N, if, in accordance with the recursive rela tion (4.18), we replace the binary numerals 1 in the (l1)th and (lp1)th digits with the binary numerals 0 and the binary numeral 0 in the lth digit with the binary numeral 1, that is, an ... al +1 0 1 al −2 ... al − p 1 al − p −2 ... a1 . N = ... a a a a a a ... 1 0 ... 0 l − p −2 1 l +1 l −2 l−p n (8.17) Such transformation of the binary representation (8.17) is named Convolu tion of the (l1)th and (lp1)th digits into the lth digit. The initial binary representation (8.17) can be restored if we fulfill the fol lowing code transformation over the binary representation of the number N: an ... al +1 1 0 al −2 ... al − p 0 al − p −2 ... a1 . N = (8.18) an ... al +1 0 1 al −2 ... al − p 1 al − p −2 ... a1 Such transformation of the code representation is named Devolution of the lth digit into the (l1)th and (lp1)th digits. Note that in accordance with (4.18) the fulfillment of the convolution and the devolution in the binary representation (8.16) does not change the initial number N, represented by the binary combination (8.16). It is simple to fulfill convolutions and devolutions for the Fibonacci pcode corresponding to the case p=1. In this case operations are carried out over three Chapter 8 Fibonacci Computers 435 adjacent digits, namely over the lth , (l1)th and (l2)th digits. Let us consider the fulfillment of these operations for the Fibonacci 1code (“Zeckendorf sum”): (a) Convolution 0 1 1 1 1 7 = 1 0 0 1 1 1 0 1 0 0 (8.19) (b) Devolution 1 0 0 0 0 5 = 0 1 1 0 0 0 1 0 1 1 (8.20) The procedure that consists in the fulfillment of all possible convolutions or devolutions in the initial binary combination (8.16) is named Code Convolution and Code Devolution, respectively. It is easy to prove that the fulfillment of the code convolution and the code devolution result in the socalled Convolute and Devolute binary representations of the number N. For the case p=1, the convolute and devolute binary representations of the number N have peculiar indications. In particular, in the convolute binary rep resentation, two binary numerals of 1 together do not meet and in the devo lute binary representation two binary numerals of 0 together do not meet since the highest 1 of the binary representation (8.16). The rule of the “convolutiondevolution inversion” is of great importance for technical applications. This rule consists in the following: the “convolution” of the initial binary combination is equivalent to the “devolution” of the inverse binary combination, and conversely. By using this rule, we can fulfill the reduc tion to the “convolute” form for the example (8.19) as follows: (a) Inversion of the initial binary combination 0 1 1 1 1: 01111 = 10000 (b) “Code devolution” of the inverse binary combination: 10000 = 01100 = 01011 (c) Inversion of the obtained code combination: 01011 = 10100 Now, let us consider the peculiarities of the “convolution” and “devolution” for the lowest digits of the Fibonacci 1code. As is wellknown in the case p=1, the weights of the two lowest digits of the Fibonacci 1code equal 1, that is, F1=F2=1. And then the operations of the “devolution” and “convolution” for these digits are fulfilled as follows: 10=01 (“devolution”) and 01=10 (“convolution”). Alexey Stakhov THE MATHEMATICS OF HARMONY 436 8.4.2. Radix of the Fibonacci pCode What is the radix of the pFibonacci code? For the case p=0, the radix of the binary system (8.10) is calculated as the ratio of the adjacent digit weights, that is, 2k/2k1=2. Apply this principle to the Fibonacci pcode (8.9) and examine the ratio (8.21) Fp(k)/Fp(k1). The radix of the Fibonacci pcode (8.9) is the limit of the ratio (8.21). From Chapter 4 we know that F (k) = τp, lim p k→∞ F (k − 1) p where τp is the golden pproportion. This means that the radix of the Fibonacci pcode (8.9) is an irrational num ber τp, the positive root of the characteristic equation xp+1xp1=0. 8.4.3. A Minimal Form of the Fibonacci pCode The following theorem is of great importance for the theory of Fibonacci pcodes. Theorem 8.3. For the given integers p≥0 and n≥p+1, the arbitrary integer N can be represented in the only form: N=Fp(n)+R1, where (8.22) 0≤R1<Fp(np). (8.23) Proof. Note that the sequence of the Fibonacci pnumbers that are generat ed by the recursive relation (4.18) at the seeds (4.19) is strictly an increasing sequence starting with n=p+1. Consider the following number sequence: (8.24) Fp(p+1), Fp(p+2),…, Fp(n), Fp(n+1),… . We can choose in this sequence the pair of adjacent Fibonacci pnumbers Fp(n) and Fp(n+1) so that (8.25) Fp(n)≤N< Fp(n+1). By subtracting the Fibonacci pnumber Fp(n) from all terms of the inequal ity (8.25), we obtain: 0≤NFp(n)<Fp(n+1)Fp(n). (8.26) The formulas (8.22) and (8.23) follow directly from (8.26) because Fp(n+1)Fp(n)=Fp(np). Chapter 8 Fibonacci Computers 437 The theorem is proved. Note that for the case p=0 we have: F0(n)=2n1 and therefore the expres sions (8.22) and (8.23) take the following wellknown (in “binary” arith metic) form: (8.27) N=2n1 +R1, 0≤R1<2n1. Let us represent the integer N according to the formula (8.22) and then let us represent all the remainders R1, R 2, Rk , which can arise in the process of such representation, according to one and the same formula (8.22). We will continue this process up to obtaining the remainder equal to 0. As a result of this decomposition of the number N, we obtain a pecu liar representation of integer N in the Fibonacci pcode (8.9). Its peculiarity is that in the binary representation of the integer N given by (8.16) no less than p binary numerals 0 follow after every digit al=1 from the left to the right, that is, (8.28) al1= al2= alp=0. Such representation of the integer N is called the Minimal Form or Minimal Representation of the integer N in the Fibonacci pcode. This name reflects the fact that for the case p=1 the minimal form of the integer N has a minimal num ber of the binary numerals 1 in the binary representation of the Fibonacci 1 code among all binary representations of the same integer N. For example, by using the above algorithm, we can obtain the following minimal forms of the number 25 in the Fibonacci 1 and 2codes (Table 8.4). Table 8.4. Minimal forms of the Fibonacci1 and 2codes p = 1 F1 (i ) 55 34 21 13 8 5 3 25 = 0 0 1 0 0 0 1 p = 2 F2 (i ) 19 13 9 6 4 3 2 25 = 1 0 0 1 0 0 0 2 1 1 0 1 0 1 1 1 0 0 0 A peculiarity of the binary representations of number 25 given by Table 8.4 consists in the following. For the case p=1, not less than one binary nu meral 0 follows after every binary numeral 1 from left to right in the binary representation of number 25; for the case p=2 not less than two binary numerals 0 follow after every binary numeral 1 from left to right in the code representa tion of the same number 25. Corollary from Theorem 8.3. For a given p (p=0,1,2,3,…) every integer N has the only minimal form in the Fibonacci pcode. This means that there is a onetoone mapping of natural numbers onto the minimal forms of the Fibonacci pcode (8.9). Alexey Stakhov THE MATHEMATICS OF HARMONY 438 Note that for the case p=0 (the classical binary code) every integer N has the only representation in the form (8.10). This means that every code representa tion (8.10) is its “minimal form.” Now, let us prove the following theorem. Theorem 8.4. For a given integer p≥0 by using the ndigit Fibonacci p code in the minimal form we can represent Fp(n+1) integers in the range from 0 to Fp(n+1)1, inclusively. Proof. To prove the theorem, recall that there is a onetoone map ping between natural numbers and their minimal forms in the Fibonacci pcode (8.9). Suppose that the number of different ndigit minimal forms of the Fibonacci pcode (8.9) is equal to T p(n). Let us find a recursive relation for Tp(n). First of all, we can write the following evidence equal ities: Tp(1)=Tp(2)=…=Tp(p)=1. (8.29) It is easy to prove that for the case n=p+1 we can represent only two num bers in the minimal form, 0 and 1, as follows: 0 = 00 ...0 ; 1 = 100 ...0 . N N p +1 p This means that Tp(p+2)=2. (8.30) Let n>p+1. Suppose that for this case the set A of ndigit minimal forms in the Fibonacci pcode (8.9) consists of Tp(n) elements, that is, we can represent Tp(n) different numbers in the minimal form. To obtain a recursive relation for Tp(n), we divide the set A into two noncrossing subsets A0 and A1. All minimal ndigit forms of the subset A0 begin from 0 and all minimal ndigit forms of the subset A1 begin from 1. By taking away the first binary numeral 0 at the begin ning of all minimal forms of the subset A0, we obtain the subset of (n1)digit minimal forms. According to the inductive hypothesis, the number of elements in it is equal to Tp(n1). Next let us examine the subset A1. In accordance with the condition for minimal form, the p binary numerals 0 follow after the binary numeral 1. This means that all binary combinations of the subset A1 begin from the code combi nation 100…0 (p+1 digits). By taking away these (p+1) digits standing at the beginning of all minimal forms of the subset A1, we obtain the set of all (np1) digit minimal forms. According to the inductive hypothesis, the number of ele ments of this subset is equal to Tp(np1). The next recursive relation for Tp(n) follows from this consideration: Tp(n)= Tp(n1)+Tp(np1). (8.31) Chapter 8 Fibonacci Computers 439 If we compare the recursive relation (8.31) that is given at the seeds (8.29) and (8.30) with the recursive relation (4.18) at the seeds (4.19), we can conclude: Tp(n)=Fp(n+1). The theorem is proved. For the case p=1, the minimal form has a very simple indicator: in the mini mal form two binary numerals 1 together do not meet. But the “convolute” form, considered above, has the same property. This means that for the Fibonacci 1 code the “convolute” form coincides with the minimal form and the reduction of the Fibonacci 1code combination to the minimal form can be performed by us ing “convolutions.” The example (8.19) demonstrates the process of reduction of the Fibonacci 1code combination to the minimal form. Also the notion of the maximal form is very important for the Fibonacci 1code. The maximal form can be obtained from the initial binary combination by means of “devolutions” and it coincides with the “devolute” form. The example (8.20) demonstrates the pro cess of reduction of the Fibonacci 1code combination from the minimal form of the number 5 to its maximal form. Note that the operations of the reduction of the Fibonacci 1code combinations to the minimal and maximal forms are the most important operations of Fibonacci arithmetic. 8.4.4. Comparison of Numbers in the Fibonacci pCodes Now, let us deduce a rule for the number comparison of Fibonacci pcodes. With this purpose we compare the ndigit numbers corresponding to the sub sets A0 and A1. As all numbers of the subset A1 have the binary numeral 1 in the highest, that is, the nth digit, the minimal number of the subset A1 is equal to Fp(n), the weight of the nth digit of the Fibonacci pcode (8.9). On the other hand, all numbers of the subset A0 have the binary numeral 0 in the highest, that is, the nth digit. Then according to Theorem 8.4, the maximal number, which may be represented in the minimal form by using (n1) digits is equal to Fp(n) 1. Hence, the minimal number of the subset A1 is always greater than the maxi mal number of the subset A0. From this examination, the simple rule of comparison follows. The compar ison of two numbers A and B represented in the minimal form of the Fibonacci pcode (8.9) can be carried out digitbydigit starting with the highest digit un til the first pair of noncoincident digits of the comparable codes A and B is found. If the number A has the binary numeral 1 and the number B has the binary numeral 0 in the first pair of the noncoincident digits, we can then conclude that A>B. If the code combinations of the comparable numbers coincide for all digits, then the numbers are equal to each other. Alexey Stakhov THE MATHEMATICS OF HARMONY 440 Thus, the comparison of numbers in the Fibonacci pcodes is carried out in a manner similar to the classical binary code, provided before the comparison we reduce the comparable codes to their minimal form. This property (simplicity of number comparison) is one of the important arithmetical advantages of the Fi bonacci pcodes (8.9). For example, we need to compare two numbers A=00111101101 and B=00111110110 that are represented in the Fibonacci 1code. The number comparison is carried out in two steps: 1. Reduce the comparable codes to their minimal form: A = 00111101101 = 0100111001 = 0101001010 B = 00111110110 = 01001111000 = 01010011000 = 01010100000. 2. Compare digitbydigit the minimal forms of the number A and B begin ning with the highest digit on the left: A = 01010 [0]10010 B = 01010 [1] 00000. We can see that the first noncoincident pair of the comparable combina tions (from left to right) contains the binary numeral 0 in the minimal form of the first number A and the binary numeral 1 in the minimal form of the second number B. This means that B>A. 8.4.5. Redundancy of the Fibonacci pCodes For the case p=0, the Fibonacci 0code (classical binary code) is non redundant. However, for p>0, all Fibonacci pcodes (8.9) are redundant. And their redundancy shows in the Multiplicity of the Fibonacci pcode represen tation of one and the same integer N. Theorems 8.2 and 8.4 allow one to calculate the redundancy of the Fibonacci pcodes for p>0 in comparison with the classical binary code (p=0). We can calculate the relative code redundancy r by the following formula [177]: n−m n r= = −1, (8.32) m m where n and m are the code length of the redundant and nonredundant codes, respectively, for the representation of one and the same number range. Note that the redundancy definition given by (8.32) characterizes a relative increase of the code length of the redundant code relative to the non redundant code for the representation of one and the same number range. Chapter 8 Fibonacci Computers 441 Now, let us examine the range of number representation for the ndigit Fi bonacci pcodes (p>0). The problem has two solutions. It follows from Theorem 8.2 that we can represent Fp(n+p+1) integers by using the ndigit Fibonacci p code (8.9) if we represent integers N≥0 without the restrictions on the form of the code representation. However, it follows from Theorem 8.4 that we can rep resent Fp(n+1) integers N by using the ndigit Fibonacci pcode (8.9) if we use only the minimal forms for the number representation. For the code representation of the numbers that are in the number range Fp(n+p+1) or Fp(n+1), it is necessary to use either m1≈log2Fp(n+p+1) or m2≈log2Fp(n+1) binary digits of the nonredundant code, respectively. Using (8.32) we obtain the following formulas for the calculation of the relative code redundancy of the Fibonacci pcode (8.9): n r1 = −1 (8.33) log 2 Fp (n + p + 1) r2 = n − 1. log 2 Fp (n + 1) (8.34) The simplest redundant Fibonacci pcode is the code corresponding to the case p=1. Let us calculate the limiting value of the relative redundancy for this code. The formulas (8.33) and (8.34) for the Fibonacci 1code take the following forms, respectively: n r1 = −1 (8.35) log 2 Fn +2 r2 = n − 1, log 2 Fn +1 (8.36) where Fn+1, Fn+2 are the classical Fibonacci numbers. We can represent the Fibonacci numbers Fn+1, Fn+2 by using the Binet for mulas (2.68). For any large n we can write the Binet formulas (2.68) in the fol lowing approximate form: τn Fn ≈ . (8.37) 5 Using (8.37) and substituting the approximate values for Fn+2 and Fn+1 into the formulas (8.35) and (8.36), we obtain the following formulas: n −1 r1 = (8.38) (n + 2)log 2 τ − log 2 5 r2 = n − 1. (n + 1)log 2 τ − log 2 5 (8.39) Alexey Stakhov THE MATHEMATICS OF HARMONY 442 If we direct n to infinity in the expressions (8.38) and (8.39), we find that they coincide for this case. Here, the utmost value of the relative redundancy for the Fibonacci 1code is determined by the following expression: 1 r= − 1 = 0.44. log 2 τ Thus, the utmost value of the redundancy of the Fibonacci 1code (Zecken dorf representation) is a constant value equal to 0.44 (44%). 8.4.6. Surprising Analogies Between the Fibonacci and Genetic Codes Among the biological concepts that have a level of general scientific signifi cance and are well formalized, the genetic code plays a special role. The discov ery of the now wellknown fact of the striking simplicity of the basic principles of the genetic code is one of the major modern discoveries of modern science. This simplicity consists of the fact that the inheritable information is encoded by the texts of the threealphabetic words Triplets or Codonums compounded on the base of the alphabet that consists of the four characters nitrogen bases: A (adenine), C (cytosine), G (guanine), T (thiamine). The given recording sys tem is unique for all boundless set of miscellaneous living organisms and is called the Genetic Code [141]. It is known [141] that by using the threealphabetic triplets or codonums, we can encode 21 items that include 20 amino acids and one more item called the Stopcodonum (sign of the punctuation). It is clear that 43=64 different combi nations (from four by three nitrogen bases) are used for encoding 21 items. In this connection some of the 21 items are encoded by several triplets. It is called the Degeneracy of the Genetic Code. Finding the conformity between triplets and amino acids (or signs of the punctuation) is interpreted as Decryption of the Genetic Code. Now, let us examine the 6digit Fibonacci 1code (Zeckendorf representation) (8.1) that uses 6 Fibonacci numbers 1, 1, 2, 3, 5, 8 as digit weights: (8.40) N = a6 × 8 + a5 × 5 + a4 × 3 + a3 × 2 + a2 × 1 + a1 × 1. There are the following surprising analogies between the 6digit Fibonacci code and the genetic code: 1. The first analogy. For the representation of numbers the 6digit binary Fibonacci code uses 26=64 binary combinations from 000000 up to 111111 which coincide with the number of triplets of the genetic code (43=64). 2. The second analogy. By using the 6digit Fibonacci code (8.40), we can represent 21 integers including the minimal number 0 that is encoded by the 6digit binary combination 000000 and the maximal number 20 that is encoded by the 6digit binary combination 111111. Note that by using the Chapter 8 Fibonacci Computers 443 triplet coding we can also represent 21 items including 20 amino acids and one additional object, the stopcodonum or the sign of the punctuation that indicates a termination of protein synthesis. 3. The third analogy. The main feature of the Fibonacci code is the Multiplic ity of number representation. Except for the minimal number 0 and the max imal number 20 that have the only code representations 000000 and 111111, respectively, all the rest of numbers from 1 up to 19 have multiple representa tions in the Fibonacci code, that is, they use no less than two code combina tions for their representation. It is necessary to note that the genetic code has the similar property called the Degeneracy of the genetic code. Thus, between the Fibonacci code (8.40) and the genetic code based on the triplet representation of amino acids, there are very interesting analogies that allow us to consider the Fibonacci code to be a peculiar class of redun dant codes among various ways of redundant coding. Thus we suggest that the study of the Fibonacci code may be of great interest for the genetic code. It is very possible that the similar analogies can become rather useful in the designing of DNAbased biocomputers. 8.5. Fibonacci Arithmetic: The Classical Approach 8.5.1. Fibonacci Addition Let us begin with developing the Fibonacci arithmetic. There are two meth ods for its development. The first way is to use an analogy between classical bi nary arithmetic and Fibonacci arithmetic (a “classical approach”). The second way is to develop the “original Fibonacci arithmetic” based on some peculiari ties of the Fibonacci pcodes. Let us first start the development with the classical approach. As the notion of the Fibonacci pcode given by (8.9) is a generalization of the notion of the classical binary code (8.10), we can use the following method for obtaining the arithmetical rules of the Fibonacci parithmetic. We begin by an alyzing the corresponding rule for the classical binary arithmetic and then by analogy we find a similar rule for the Fibonacci parithmetic. We start from the addition rule for the Fibonacci pcode (8.9). It is well known that the classical binary addition is based on the following identity for the binary numbers: 2 k1+2k1=2k, (8.41) Alexey Stakhov THE MATHEMATICS OF HARMONY 444 where 2k1 and 2k are the weights of the (k1)th and kth digits of the binary code (8.10), respectively. To obtain the rule of number addition for the Fibonacci pcode, we ana lyze the following sum: (8.42) Fp(k)+Fp(k), where Fp(k) is the weight of the kth digit of the Fibonacci pcode (8.9). Let p=1. For this case we have (8.43) Fp(k)=Fk, where Fk is the classical Fibonacci number given by the recursive relation (8.2) at the seeds (8.3). If we take into consideration (8.2) and (8.3), we can represent the sum (8.42) as follows: Fk+Fk=Fk+Fk1+Fk2 (8.44) Fk+Fk=Fk+1+Fk2. (8.45) The following addition rule of two binary digits ak + bk with the same index k for the Fibonacci 1code follows from (8.44) and (8.45) represented in Table 8.5. The following rules of the Fibonacci 1addition of the binary numerals of the kth digits ak+bk follow from Table 8.5. A rule of summation Table 8.5: for the Fibonacci 1code Rule 8.1. At the addition of ak + bk = ck +1 sk ck −1 ck −2 the binary numerals 1+1 of the k 0 + 0 = 0 0 0 0 th digits the carryover of the bi 0 + 1 = 0 1 0 0 nary numeral 1 from the kth dig 1 + 0 = 0 1 0 0 it to the next two digits arises. Rule 8.2. There are two 1 + 1 = 0 1 1 1 (a) 0 0 1 (b) ways of carryingover the for 1 + 1 = 1 mation. In method (a) (see Ta ble 8.5) the binary numeral 1 is attributed to the kth digit of the intermediate sum sk and the carryover of the binary numeral 1 to the next two lower digits, that is, to the (k1)th and (k2)th digits, ck1 and ck2, occurs. Method (b) (see Table 8.5) assumes another rule of the Fibonacci 1summation of binary dig its. The binary numeral 0 is attributed to the kth digit of the intermediate sum sk and the binary numeral 1 is carriedover to the next two digits, that is, to the (k+1)th and (k2)th digits, ck+1 and ck2, appear. The summation of the multidigit numbers in the Fibonacci 1code is fulfilled in accordance with the Fibonacci 1summation according to Table 8.5. However, we should adopt the following rules: Rule 8.3. Before summation, the addends are reduced to the minimal form. Chapter 8 Fibonacci Computers 445 Rule 8.4. In accordance with Table 8.5, it is necessary to form the multi digit intermediate sum and multidigit carryover. Rule 8.5. The multidigit intermediate sum is reduced to the minimal form and is then summarized with the multidigit carryover. Rule 8.6. The addition process continues in accordance with rules 8.4 and 8.5 until the multidigit carryover equal to 0 is obtained. The last intermediate sum, which is reduced to the minimal form, is the result of the addition. For the Fibonacci 1addition it is necessary to fulfill the additional rule 8.7. Rule 8.7. Let us examine the case, where we have two binary numerals 1 in the kth digits of the addends. It follows from the property of the minimal form that the binary numerals of the (k+1)th and (k1)th digits of both of the addends are always equal to 0. It is clear that for this case the intermediate sum that appears at the addition of the (k+1)th and (k1)th digits of both of the addends are also always equal to 0. This means that we can place one of the carryovers that appear at the addition of the kth significant digits (1+1) at once to the (k1)th digit of the intermediate sum (for the (a)method) or to the (k+1)th digit of the intermediate sum (for the (b)method). We can demonstrate the above Fibonacci addition rules in the example below. Example 8.1. Summarize the numbers 31=10011011 and 22=01011010 represented in the Fibonacci 1code. The first step is to reduce the binary representations of the numbers 31 and 22 to the minimal form: 31=10100100; 22=10000010. The second step is the formation of the multidigit intermediate sum S1 and multidigit carryover C1 in accordance with method (a) of Table 8.5: 31 = 10100100 + 22 = 10000010 ___________ S1 = 11100110 C1 = 00100000 The third step is to reduce the intermediate sum S1 to the minimal form: S1=11100110=100101000. The fourth step is the addition of S1 and C1: S1 = 100101000 + C1 = 000100000 ____________ S2 = 100111000 C2 = 000001000 Alexey Stakhov THE MATHEMATICS OF HARMONY 446 The fifth step is to reduce the intermediate sum S2 to the minimal form: S2=100111000=101001000. The sixth step is the addition of S2 and C2: S2 = 101001000 + C2 = 000001000 ____________ S3 = 101001100 C3 = 000000010 The seventh step is to reduce the intermediate sum S3 to the minimal form: S3=101001100=101010000. The eighth step is the addition of S3 and C3: S3 = 101010000 + C3 = 000000010 ____________ S4 = 101010010 C4 = 000000000 The addition is over because the carryover C4=0 appears in the eighth step. 8.5.2. Direct Fibonacci Subtraction The wellknown “direct” method of number subtraction in classical bina ry arithmetic (p=0) is based on the following property of binary numbers: (8.46) 2n+k2n=2n+k1+2n+k2+…+2n . Now, let us write a similar identity for the Fibonacci 1numbers: (8.47) Fn+k Fn= Fn+k2+Fn+k3+…+Fn1 . Using the identity (8.47) and the Fibonacci recursive relation (8.2), we obtain the following Fibonacci 1subtraction table represented in Table 8.6. The reader of the present book can “deduce” the rules of Table 8.6. The the “direct” Fibonacci psubtraction by using the recursive rule of subtrac relation for the Fibonacci pnumbers. tion for the Fibonacci 1code The “direct” Fibonacci 1subtraction of multidigit num 0−0=0 bers is based on the following rules: 1−1=0 Rule 8.8. The numbers are reduced to the minimal form 1 − 0 = 0 1 1 before subtraction. 10−1=01 Rule 8.9. After reduction to the minimal form, the numbers 100−1=11 are compared by their value and then in accordance with Table 1 0 0 0 − 1 = 1 1 1 8.6, the smaller number being subtracted from the larger one. Chapter 8 Fibonacci Computers 447 8.5.3. Fibonacci Inverse and Additional Codes It is wellknown that the classical binary arithmetic employs the notions of “inverse” and “additional” codes for the representation of negative numbers. This approach allows one to reduce the binary subtraction to binary addition. For the case of integer representation, the following mathematical properties under lie the “inverse” code (Ginv) and the “additional” code (Gad): Ginv=2n1 |G| (8.48) Gad=2n |G|, (8.49) where |G| is the module of the initial integer G. Note that the “” (“minus”) sign is encoded by the binary numeral 1 and the “+” (“plus”) sign by the binary numeral 0. They are placed at the begin ning of the code combination. The digital part coincides with its direct code for the positive number G, and is determined by (8.48) and (8.49) for the negative number G. The notions of inverse and additional codes have the following generaliza tion for the Fibonacci 1code. By analogy to (8.48) and (8.49) we can introduce the notions of the “Fibonacci inverse” and “Fibonacci additional” ndigit 1codes: Ginv=Fn+11 |G| (8.50) Gad=Fn+1 |G|, (8.51) where Fn+1 is the Fibonacci 1number, and G is any integer. The expressions (8.50) and (8.51) have the following code interpretation. Suppose that the initial integer G is represented in the minimal form of the n digit Fibonacci 1code (8.1). It follows from Theorem 8.4 that the integer G is in the following number range: (8.52) 0 ≤ G ≤ Fn −1. On the other hand, the following important expressions that connect the Fibonacci inverse and additional codes follow from (8.50) and (8.51): Ginv + G = Fn −1 − 1 (8.53) (8.54) Gad = Ginv + 1. Using the general property (2.20), we can represent the number Fn+1–1 as follows: Fn+1–1= Fn1+Fn2+…+F2+F1. Note that the sum (8.55) has the following code interpretation: Fn +1 − 1 = 011 ...1 = Ginv + G . N n −1 (8.55) (8.56) Alexey Stakhov THE MATHEMATICS OF HARMONY 448 The formula (8.56) can be used for the introduction of the following algo rithm for obtaining the Fibonacci inverse and Fibonacci additional 1codes from the initial code G (for example, from the initial Fibonacci 1code G=10010100): 1. By using the operation of “devolution” for the initial code G: G=10010100=01110011=01101111. 2. By using the operation of “inversion” for all binary digits of the “devolute” form of the number G, except for the higher digit: 01101111 = 00010000. The resulting binary combination is the Fibonacci inverse 1code of G: Ginv=00010000. 3. By adding the binary numeral 1 to the lowest digit of the Fibonacci in verse 1code in accordance to (8.54) and by reducing the obtained code com bination to the minimal form, we obtain the Fibonacci additional 1code G: Gad=00010001=00010010. Note that in the general case p≥0, the expressions for the Fibonacci inverse and additional pcodes have the following forms: Ginv = Fp ( n + 1) − 1 − G (8.57) Gad = Fp ( n + 1) − G . (8.58) We can see that the expressions (8.48), (8.49) and (8.53), (8.54) are partial cases of expressions (8.57) and (8.58) because for the case p=0, F0(n+1)=2n and for the case p=1, F1(n+1)=Fn+1. Now, let us introduce the following rule of the sign encoding for the Fibonacci inverse and additional 1codes. We will be encoding the “−” sign with the binary numeral 1 and the “+” sign with the binary numeral 0. The sign of the code com binations is put at the beginning of the numerical part of the code. Hence, the Fibonacci inverse and additional 1codes of the negative number G=10010100 have the following forms, respectively: Ginv=1.00010000 Ginv=1.00010010. Note that the numerical part of the Fibonacci inverse and additional 1 codes of the positive number G, coincide with the Fibonacci 1code of the same number G. For example, the Fibonacci inverse and additional 1codes of the positive number G=+10010100 have the following form: Gad=0.10010100. Chapter 8 Fibonacci Computers 449 Let us formulate the Fibonacci subtraction algorithm based on the notions of the Fibonacci inverse and additional 1codes: 1. Represent the initial numbers in the Fibonacci inverse or additional 1 codes. 2. Sum the numerical parts in accordance with the rule of the Fibonacci 1 addition and the signs in accordance with the rule of classical binary addi tion. The addition result is represented in the Fibonacci inverse or addition al 1codes. If the sign of the code combination is equal to 1, it means that the addition result is a negative number. For obtaining the absolute value of the addition result, it is necessary to use the above algorithm of code transfor mation for the Fibonacci inverse or additional 1codes for the numerical part of the addition result. 8.6. Fibonacci Arithmetic: An Original Approach 8.6.1. Basic Micro Operations As mentioned above the multiplicity of number representation is the main peculiarity of the Fibonacci pcode (8.9) for the case p>0 in compar ison to the classical binary code (8.10). By using the above operations of “devolution” and “convolution”, we can change the form of representation of one and the same number. This means that the binary 1’s can move in the code combination to the left or to the right. This fact allows us to develop an original approach to the Fibonacci arithmetic based upon the socalled Basic Microoperations. We can introduce the following four “basic microoperations” that are used in the Fibonacci processor for fulfillment of the logical and arithmetical opera tions: (a) Convolution; (b) Devolution; (c) Replacement; (d) Absorption. Consider these microoperations for the case of the Fibonacci 1code (Zeckendorf representation). We showed above that for the case p=1 the Convolution and Devolution are the following code transformations that are carried out for all the adjacent triple digits of one and the same Fibonacci 1 code combination. Convolution: [011→100] Devolution: [100→011] Alexey Stakhov THE MATHEMATICS OF HARMONY 450 The microoperation of Replacement 1 0 ↓ = 0 1 is a twoplaced microoperation that is fulfilled at the same digit of two reg isters, the top register A and the bottom register B. Examine the case when the register A has the numeral 1 in the kth digit and the register B has the numeral 0 in the same digit. For this condition we can fulfill the following microoperation. We shift the above binary numeral 1 of the top register A to the bottom register B. This microoperation is named Replacement. Note that this operation can be carried out only for the condition if the kth digits of the registers A and B are equal to 1 and 0, respectively. The microoperation of Absorption 1 0 7 = 1 0 is a twoplaced microoperation that consists in mutually annihilating two 1’s in the kth digit of two registers A and B and replacing them with the binary numerals 0. It is necessary to pay attention to the following “technical” peculiarity of the above “basic microoperations.” At register interpretation of these microopera tions, each microoperation may be considered to be the inversion of the triggers (flipflops) that are involved in the microoperation. This means that each mi crooperation is carried out by trigger switching. 8.6.2. Logical Operations We can demonstrate the possibility of carrying out all logical operations by using the above four “basic microoperations.” Let us construct all the possible “replacements” from top register A to bottom register B: A = 1 0 0 1 0 0 1 ↓ ↓ B = 0 1 0 0 1 0 1 A′ = 0 0 0 0 0 0 1 B′ = 1 1 0 1 1 0 1 We obtain two new code combinations A′ and B′ as the outcome of the “re placement.” We can see that the binary combination A′ is a logical “conjunction” (∧) of the initial binary combinations A and B, that is, Chapter 8 Fibonacci Computers 451 A′ = A ∧ B and the binary combination B′ is a logical “disjunction” (∨) of the initial code combinations A and B , that is, B′ = A ∨ B . The logical operation of “module 2 addition” is performed by means of the simultaneous fulfillment of all possible microoperations of “replacement” and “absorption.” For example, A = 1 0 1 0 0 1 1 0 1 7 ↓ ↓ ↓ 7 B = 1 1 0 0 1 0 0 0 1 A′ = 0 0 0 0 0 0 0 0 0 = const 0 B′ = 0 1 1 0 1 1 1 0 0 = A ⊕ B We can see that the two new code combinations A′ = const 0 and B′ = A ⊕ B are the result of this code transformation. The logical operation of “code A inver sion” is reduced to fulfillment of the operation of “absorption” at the initial code combination A and the special binary combination B=const 1 (see below): A = 1 0 1 0 0 1 1 0 1 7 7 7 7 7 B = 1 1 1 1 1 1 1 1 1 A′ = 0 0 0 0 0 0 0 0 0 = const 0 B′ = 0 1 0 1 1 0 0 1 0 = A 8.6.3. The Counting and Subtracting of Binary 1 Let us demonstrate the possibility of carrying out the simplest arithmet ical operations by using the “basic microoperations.” We start with the oper ations of “counting” and “subtracting” of the binary numerals 1. The “counting” of the binary numerals 1 in the Fibonacci 1code (a “summing counter”) is carried out by using “convolution.” For example, the transformation of the initial Fibonacci 1code combination 01010 the minimal form of the number 4 Fi = 5 3 2 1 1 4 = 0 1 0 1 0 to the next Fibonacci code combination of the number 5 is carried out in the “summing Fibonacci counter” according to the rule: Alexey Stakhov THE MATHEMATICS OF HARMONY 452 Fi = 5 3 2 1 1 4 = 0 1 0 1 0 + 1 5 = 0 1 0 1 1 5 = 0 1 1 0 0 5 = 1 0 0 0 0 Here, in the top row we see Fibonacci numbers 5, 3, 2, 1, 1 that are the weights of the Fibonacci 1code. The second row is a representation of the number 4 in the minimal form (01010). We can see that the binary numeral 1 is added (+1) to the lowest digit of the Fibonacci binary combination 01010. Then, the initial binary combination 4=01010 is transformed into the binary combination 5=01011 that is the Fibonacci binary representation of the next number 5 (the third row). After that the Fibonacci binary combination 01011 is reduced to the minimal form. This transformation is fulfilled in 2 steps by using the “convolutions.” The first step is to carry out the “convolution” over the three lowest digits of the binary combination 01011 (the third row). Using the “convolution” [011→100], we transform the Fibonacci binary representa tion of the number 5=01011 to another Fibonacci binary representation of the same number 5=01100 (the fourth row). Then, we carry out the “convolution” over the next group 011 of the Fibonacci binary combination 5=01100 (the fourth row). In the fifth row we see the minimal form of the number 5=10000. We can continue the “counting” of the binary numerals 1 as follows: 6 = 10000 + 1 = 10001 = 10010 7 = 10010 + 1 = 10011 = 10100. If we add the binary numeral 1 to the lowest digit of the number 7=10100, we obtain: 10100+1=10101=10110=11000=00000. This situation is well known in computer engineering and is named “over filling” of the “summing counter.” The “subtracting” of the binary numerals 1 (a “subtracting counter”) is fulfilled by using “devolution.” For this purpose the initial Fibonacci 1code combinations are reduced to the “devolute” or “maximal” form, and then the binary numeral 1 is subtracted sequentially from the lowest digit: Fi = 5 3 2 1 1 5 = 1 0 0 0 0 5 = 0 1 1 5 = 0 1 0 1 1 − 1 4 = 0 1 0 1 0 Chapter 8 Fibonacci Computers 453 Here, in the second row we see the representation of the number 5 in the minimal form (5=10000). In the third row we carry out the “devolution” over the highest three digits [100→011] of the binary representation of the initial binary combination 5=10000. In the fourth row we carry out the next “devolu tion” over the next three digits of the binary representation of the number 5=01100 (see the third row). As the outcome we obtain the maximal form of the number 5=01011 (the fourth row). Then, we subtract the binary numeral 1 from the lowest digit of the binary combination 5=01011. A result of this transforma tion 4=01010 is represented in the fifth row. Then, we can continue the “sub tracting” of the 1’s as follows: 4 = 01010 = 01001 (" devolutions ") 3 = 4 − 1 = 01001 − 1 = 01000 3 = 01000 = 00110 = 00101 (" devolutions ") 2 = 3 − 1 = 00101 − 1 = 00100 2 = 00100 = 00011 (" deevolution ") 1 = 2 − 1 = 00011 − 1 = 00010 1 = 00010 = 00001 (" devolution ") 0 = 1 − 1 = 00001 − 1 = 00000. 8.6.4. Fibonacci Summation The idea of summation of two numbers A and B is based on the “basic micro operations.” The first step is to shift all binary numerals 1 from the top register A to the bottom register B. With this purpose we use “replacement,” “devolution,” and “convolution.” The sum is formed in register B. For example, let us sum the following numbers A0=010100100 and B0=001010100. The first step of the Fibonacci addition consists in the “replacement” of all possible binary numerals 1 from register A to register B: A0 = 0 1 0 1 0 0 1 0 0 ↓ ↓ B0 = 0 0 1 0 1 0 1 0 0 A1 = 0 0 0 0 0 0 1 0 0 B1 = 0 1 1 1 1 0 1 0 0 The second step is the fulfillment of all possible “devolutions” in the binary combination A1 and all possible “convolutions” in the binary combination B1: Alexey Stakhov THE MATHEMATICS OF HARMONY 454 A1=000000100→ A2=000000011 B1=011110100→ B2=100110100. The third step is the “replacement” of all possible binary numerals 1 from register A to register B: A2 = 0 0 0 0 0 0 0 1 1 ↓ ↓ B2 = 1 0 0 1 1 0 1 0 0 A3 = 0 0 0 0 0 0 0 0 0 B3 = 1 0 0 1 1 0 1 1 1 The addition is over because all binary numerals 1 are shifted from the regis ter A to the register B. After reducing the binary combination B3 to the minimal form, we obtain the sum B3=A+B, which is represented in the minimal form: B3=100110111=101001001=101001010=A+B. Thus, the addition is reduced to a sequential fulfillment of the microopera tions of “replacement” over two binary combinations A and B and the micro operations of both “convolution” over the binary combination B and “devolu tion” over the binary combination A. 8.6.5. Fibonacci Subtraction The idea of direct subtraction of number B from number A which is based on the “basic microoperations,” consists of the mutual “absorption” of the binary numerals 1 in the binary combinations of numbers A and B until one of them becomes equal to 0. To perform this process, we have to sequentially carry out the aforementioned microoperations of “absorption” and “devolution” over the code combinations A and B. The subtraction result is always formed in the register that contains the larger number. If the subtraction result is formed in the top register A, it follows from this fact that the sign of the subtraction result is “+” (“plus”); in the opposite case the subtraction result has the “−” (“minus”) sign. Let us now demonstrate this idea for the following example. Subtract the number B0=101010010 from the number A0=101001000 in the Fibonacci 1code. The first step is “absorption” of all possible binary 1’s in the numbers A and B: A0 = 1 0 1 0 0 1 0 0 0 7 7 B0 = 1 0 1 0 1 0 0 1 0 A1 = 0 0 0 0 0 1 0 0 0 B1 = 0 0 0 0 1 0 0 1 0 Chapter 8 Fibonacci Computers 455 The second step is “devolution” of the code combinations A1 and B1: A1=000001000→ A 2=000000110 B1=000010010→ B2=000001101. The third step is “absorption” of A2 and B2: A2 = 0 0 0 0 0 0 1 1 0 7 B2 = 0 0 0 0 0 1 1 0 1 A3 = 0 0 0 0 0 0 0 1 0 B3 = 0 0 0 0 0 1 0 0 1 The fourth step is “devolution” of A3 and B3: A3=000000010→ A4=000000001 B3=000001001→ B4=000000111. The fifth step is “absorption” of A4 and B4: A4 = 0 0 0 0 0 0 0 0 1 7 B4 = 0 0 0 0 0 0 1 1 1 A5 = 0 0 0 0 0 0 0 0 0 B5 = 0 0 0 0 0 0 1 1 0 The subtraction is finished. After reducing the code combination B5 to the minimal form we obtain: B5=000001000. The subtraction result is in the register B. This means that the sign of the subtraction result is “−” (“minus”), that is, the difference of the numbers AB is equal to: R=AB=000001000. 8.7. Fibonacci Multiplication and Division 8.7.1. The Egyptian Method of Multiplication In order to find the algorithms for Fibonacci multiplication and division, we use an analogy with classical binary multiplication and division. We start with multiplication. To multiply two numbers A and B in the classical binary code Alexey Stakhov THE MATHEMATICS OF HARMONY 456 (8.10) we should represent the multiplier B in the form of the ndigit binary code. The product P=A×B can then be written in the following form: (8.59) P=A×bn2n1+ A×bn12n2+…+ A×bi2i1+…+ A×b120, where bi is the binary numeral of the multiplier B. It follows from (8.59) that the binary multiplication is reduced to form the partial products of the kind A×bi2i1 and their addition. The partial product A×bi2i1 is made up by shifting the binary code of number A to the left by (i1) digits. The binary multiplication algorithm based on (8.59) has a long history and goes back in its origin to the Egyptian Doubling Method [169]. As is known, the Egyptian number system was decimal, but nonposition al. The Egyptians represented the “nodal” numbers of their number system 1, 10, 100, 1000, 10 000, etc. by using special hieroglyphs. Then, in the record of any number, for example, 325, they used the 5 hieroglyphs representing the “nodal” number 1, the 2 hieroglyphs representing the “nodal” number 10, and the 3 hieroglyphs representing the “nodal” number 100. The invention of the Doubling Method that underlies the arithmetical oper ations of multiplication and division is the main achievement of Egyptian arith metic. In order to multiply the number 35 by the number Table 8.7. Egyptian 12, the Egyptian mathematician acted as follows. He built method of multiplication up a special table of numbers (see Table 8.7). The binary numbers 2k (k=0,1,2,3,…) were placed in the first column 1 35 of Table 8.7. In the second column of Table 8.7 we can see 2 70 the numbers 35, 70, 140, 280, that is, the numbers formed /4 140 → 140 from the multiplier 35 according to the Doubling Method. /8 280 → 280 420 After that the Egyptian mathematician used the in clined line to mark those binary numbers of the first column whose sum is equal to the second multiplier (12=8+4). Then he selected those numbers of the second column corresponding to the marked “binary” numbers of the first column. The result of multiplication is equal to the sum of the selected numbers of the second column (140+280=420). The analysis of the Egyptian method of multiplication based on the Dou bling Method results in a rather unexpected conclusion. The representation of the multiplier 12 in the form 12=8+4 corresponds to its representation in the binary system (12=1100). On the other hand, if we represent the second multiplier 35 in the binary system (35=100011), then the “doubling” of the number 35 corresponds to the displacement of the binary code of the multi plier 35=100011 to the left (70=1000110, 140=10001100, etc.). In other words, the Egyptian multiplication method based on the Doubling Method coincides with the basic multiplication algorithm used in modern computers! Chapter 8 Fibonacci Computers 457 This basic algorithm follows from the representation of the product in the form of (8.56). 8.7.2. The Egyptian Method of Division The Egyptians also used the Doubling Method for the division of numbers. If, for example, we need to divide the number 30 by the number 6, the Egyp tian mathematician acted as follows (see Table 8.8). The first stage of the division consists of the following. Let us write in the first column of Table 8.8 the binary numbers 2k (k=0, 1, 2, 3,…). Then, let us write in the second column the divisor 6 and Table 8.8. Egyptian method its “doubling” numbers 12, 24, and 48. In ev of division ery “doubling” step we will compare the divi 6 ≤ 30 6≤6 /1 dend 30 with the numbers of the second col 2 12 ≤ 30 12 > 6 umn until the next “doubling” number (the 24 ≤ 30 /4 number 48 in this example) becomes more 8 48 > 30 than the dividend (48>30). After that we shall 4 + 1 = 5 30 − 24 = 6 6 − 6 = 0 subtract the “doubling” number of the previ ous row (3024=6) from the dividend 30 and mark the corresponding binary number of the first column (the number 4) with the inclined line. After that we shall perform the same procedure with the remainder 6 (see the third column) until the “doubling” number becomes strictly more than the differ ence 6 (12>6). By marking the number 1 of the first column with the inclined line and then by subtracting the divisor 6 from the remainder 6, we obtain the number 0 (66=0). This means that the division is over. After that we can see that the sum of the binary numbers of the first column marked by the inclined line (4+1=5) is equal to the result of the division. Thus, after due consideration one is astonished by the genius of the Egyp tian mathematicians who several millennia ago invented the methods of mul tiplication and division that are today used in our modern computers! 8.7.3. Fibonacci Multiplication Analysis of the Egyptian Doubling Method allows us to develop the follow ing method of “Fibonacci pmultiplication.” Let us consider the product P=A×B, where the numbers A and B are rep resented in the Fibonacci pcode (8.9). By using the representation of the multiplier B in the Fibonacci pcode (8.9), we can write the product P=A×B in the following form: Alexey Stakhov THE MATHEMATICS OF HARMONY 458 P=A×bnFp(n)+A×bn1Fp(n1)+…+A×bi Fp(i)+…+A×b1 Fp(1), (8.60) where Fp(i) is the Fibonacci pnumber. Note that the expression (8.60) is a generalization of the expression (8.59) that underlies the algorithm of the “binary” multiplication. The fol lowing algorithm of the Fibonacci pmultiplication follows directly from the expression (8.60). The multiplication is reduced to the addition of the partial products of the kind A×biFp(i). They are formed from the multiplier A according to the special procedure that reminds one of the Egyptian Dou bling Method. Let us demonstrate the “Fibonacci multiplication” for the case of the simplest Fibonacci 1code (p=1). Example 8.2. Find the following product: 41×305. Solution: 1. Construct a table consisting of Table 8.9. Fibonacci multiplication three columns marked by F, G and P F G P (see Table 8.9). 1 305 2. Insert the Fibonacci 1sequence 1 305 (the classical Fibonacci numbers) 1, /2 610 → 610 1, 2, 3, 5, 8, 13, 21, 34 into the Fcol 3 915 umn of Table 8.9. /5 1525 → 1525 3. Insert the generalized Fibonacci 1 8 2440 sequence: 305, 305, 610, 915, 1525, 13 3965 2440, 3965, 6505, 10370, which is built 21 6505 up from the first multiplier 305 ac /34 10370 → 10370 cording to the “Fibonacci recursive 41 = 34 + 5 + 2 41×305 = 12505 relation” into the Gcolumn. 4. Mark by inclined line (/) and bold type all numbers of the Fcolumn, which make up the second multiplier 41 in the sum (41=34+5+2). 5. Mark by bold type all Gnumbers corresponding to the marked Fnum bers and rewrite them in the Pcolumn. 6. Summing all Pnumbers, we obtain the product: 41×305=12505. The multiplication algorithm is easily generalized for the case of Fibonac ci pcodes. 8.7.5. Fibonacci Division We can apply the above Egyptian method of division to construct the algorithm of Fibonacci pdivision. Consider this method for the following specific example. Chapter 8 Fibonacci Computers 459 Example 8.3. Divide the number 481 (the dividend) by the number 13 (the divisor) in the Fibonacci 1code. Solution has two steps as shown below. 8.7.5.1. The first stage 1. Construct the table, which consists of three columns marked by F, G and D (see Table 8.10). 2. Insert the Fibonacci 1sequence (the classical Fibonacci numbers) 1, 1, 2, 3, 5, 8, 13, 21, 34, 55 to the Fcolumn of Table 8.10. 3. Insert the generalized Fibonacci 1se Table 8.10. The first stage of the Fibonacci division quence: 13, 13, 26, 39, 65, 104, 169, 273, F G D 442, 615 formed from the divisor 13 ac ≤ 1 13 481 cording to the “Fibonacci recursive rela ≤ 1 13 481 tion” into the Gcolumn. ≤ 2 26 481 4. Compare sequentially every Gnumber ≤ 3 39 481 with the dividend 481, inscribed into the ≤ 5 65 481 Dcolumn, and fix the result of compari ≤ 8 104 481 son (≤ or >) until we obtain the first com ≤ 13 169 481 parison result of the kind: 615>481. ≤ 481 21 273 5. Mark with the inclined line (/) and bold /34 442 ≤ 481 type the Fibonacci number 34, which cor 55 615 > 481 responds to the preceding Gnumber 442, R1 = 481− 442 = 39 and mark the latter with bold type. 6. Calculate the difference: R 1=481 442=39. Table 8.11. The second stage of the Fibonacci division F 1 1 2 /3 5 R2 G D 13 39 ≤ 13 39 ≤ 26 39 ≤ 39 39 ≤ 65 > 39 = 39 − 39 = 0 8.7.5.2. The second stage The second stage of the Fibonacci 1division is a repetition of the first stage when we instead use the dividend 481, the remainder R1=39 (see Table 8.11). The second remainder R 2=39–39=0, which means the Fibonacci 1division is complete. The result of the division is equal to the sum of all the marked Fnumbers obtained throughout all stages (see Tables 8.10 and 8.11), that is, 34+3=37. Alexey Stakhov THE MATHEMATICS OF HARMONY 460 8.8. Hardware Realization of the Fibonacci Processor 8.8.1. A Device for Reduction of the Fibonacci Code to Minimal Form The devices for the “convolution” and “devolution” play an important role in the technical realization of arithmetical operations in the Fibonacci code. They can be designed on the basis of the binary register, which has special logical circuits for the “convolutions” and “devolutions.” Each digit of the register contains a binary “flipflop” (trigger) and logical elements. The operations of “convolution” [011→100] and “devolution” [100→011] can be carried out by means of the inversion of the “flipflops” (triggers). One of the possible variants of the “convolution” register or “the device for the reduction of the Fibonacci code to the minimal form” is shown in Fig. 8.1. This device consists of the five RStriggers and the logical elements AND, OR which are used to carry out the “convolution.” The “convolution” is carried out by means of the logical elements AND1 AND5 and OR which stand before the R and Sinputs of the triggers. The logical element AND1 fulfills the “convolution” of the 1st digit to the 2nd digit. Its two inputs are connect ed with the direct output of trigger T1 and the inverse output of trigger T2. The 3rd input is connected with the synchronization input C. The logical element AND1 analyzes the states Q1 and Q2 of the triggers T1 and T2. If Q1=1 and Q2=0, it means that the convolution conditio